# Multiple cylindrical source as shown in figure **URL:** \ **Category:** Source Definition\ **Created:** [18 February 2022 04:26 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806 "2022-02-18T04:26:53Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![ArghyaC](https://fluka-forum.web.cern.ch/user_avatar/fluka-forum.web.cern.ch/arghyac/32/46_2.png) [@ArghyaC](https://fluka-forum.web.cern.ch/u/ArghyaC)\ **Post date:** [18 February 2022 04:26 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/1 "2022-02-18T04:26:54Z")
Dear Experts, Is it possible to model multiple cylindrical sources (as shown in the attached figure) in FLUKA ? How to sample different orientation of the source cylinders in a single source.f ? Can these cylindrical sources be activated in region-basis option of newgen\_source.f as shown below? call regionsource ( coordinate\_x, coordinate\_y, coordinate\_z, [region number], [xmin], [xmax], [ymin], [ymax], [zmin], [zmax] ) Here how to put the [region number] of these many cylindrical source regions? Waiting for your suggestions. Thanking you in advance. [design\_Silver hybrid eye plaque I-125 capsule.pdf](https://fluka-forum.web.cern.ch/uploads/short-url/jeewhKXoxbHkzX7T4qR3tLbgXUL.pdf) (43.7 KB) ---
**Author:** ![horvathd](https://fluka-forum.web.cern.ch/user_avatar/fluka-forum.web.cern.ch/horvathd/32/10584_2.png) [@horvathd](https://fluka-forum.web.cern.ch/u/horvathd)\ **Post date:** [18 February 2022 09:52 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/2 "2022-02-18T09:52:34Z")
Dear @ArghyaC, Similarly as it was discussed in [Multiple complex source in Fluka](https://fluka-forum.web.cern.ch/t/multiple-complex-source-in-fluka/1241), you can use a random number to select between regions. Please note, the `regionsource()` subroutine is not part of the distributed version of the _source\_newgen.f_ user routine. Also, since it is using a rejection method, the speed of the sampling is worse, than sampling directly from a cylinder and then roto-translating the coordinates to match the geometry. Cheers, David ---
**Author:** ![ArghyaC](https://fluka-forum.web.cern.ch/user_avatar/fluka-forum.web.cern.ch/arghyac/32/46_2.png) [@ArghyaC](https://fluka-forum.web.cern.ch/u/ArghyaC)\ **Post date:** [18 February 2022 10:55 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/3 "2022-02-18T10:55:26Z")
Thank you for quick reply. Can I code like below: * * * random = FLRNDM(XDUMMY) if (random .lt. beam1\_probablility) then call regionsource ( coordinate\_x, coordinate\_y, coordinate\_z, [region number\_1], [xmin], [xmax],[ymin], [ymax], [zmin], [zmax] ) else if (random .lt. beam1\_probability + beam2\_probability) then call regionsource ( coordinate\_x, coordinate\_y, coordinate\_z, [region number\_2], [xmin], [xmax], [ymin], [ymax], [zmin], [zmax] ) else if (random .lt. beam1\_probability + beam2\_probability + beam3\_probability) then call regionsource ( coordinate\_x, coordinate\_y, coordinate\_z, [region number\_3], [xmin], [xmax], [ymin], [ymax], [zmin], [zmax] ) else call regionsource ( coordinate\_x, coordinate\_y, coordinate\_z, [region number\_4], [xmin], [xmax], [ymin], [ymax], [zmin], [zmax] ) end if * * * Kindly correct me if the above code is wrong. Whether you have some example source.f file using roto-translating the coordinates for such application? Kindly give me some suggestions. ---
**Author:** ![horvathd](https://fluka-forum.web.cern.ch/user_avatar/fluka-forum.web.cern.ch/horvathd/32/10584_2.png) [@horvathd](https://fluka-forum.web.cern.ch/u/horvathd)\ **Post date:** [18 February 2022 11:03 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/4 "2022-02-18T11:03:30Z")
Dear @ArghyaC, The code looks correct, but you need to optimize the x/y/z min/max values for each regions separately, to increase the success rate of the rejection sampling. I’m not aware of an example for a roto-translation, but since it is only 2 dimensional matrix and vector operations are needed, it should be hard to program them. ([Rotation matrix - Wikipedia](https://en.wikipedia.org/wiki/Rotation_matrix)) Cheers, David ---
**Author:** ![ArghyaC](https://fluka-forum.web.cern.ch/user_avatar/fluka-forum.web.cern.ch/arghyac/32/46_2.png) [@ArghyaC](https://fluka-forum.web.cern.ch/u/ArghyaC)\ **Post date:** [18 February 2022 11:22 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/5 "2022-02-18T11:22:25Z")
Thank you for your reply and co-operations. Since it is a cylindrical region, can I use ([Rmin], [Rmax], [Zmin], [Zmax]) instead of ([xmin], [xmax], [ymin], [ymax], [zmin], [zmax] ) for calling region source ? ---
**Author:** ![horvathd](https://fluka-forum.web.cern.ch/user_avatar/fluka-forum.web.cern.ch/horvathd/32/10584_2.png) [@horvathd](https://fluka-forum.web.cern.ch/u/horvathd)\ **Post date:** [18 February 2022 12:00 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/6 "2022-02-18T12:00:37Z")
Dear @ArghyaC, no, you can’t change the arguments unless you alter the subroutine accordingly. The x/y/z min/max values set up a bounding box around the desired region. A point is uniformly sampled in this box, and checked if it is inside the region. If not, a new point is sampled. So if you change it to a cylinder, you will encounter the problem of the rotated regions, as they won’t align with the bounding region. Cheers, David ---
**Author:** ![rversaci](https://fluka-forum.web.cern.ch/letter_avatar_proxy/v4/letter/r/cc9497/32.png) [@rversaci](https://fluka-forum.web.cern.ch/u/rversaci)\ **Post date:** [20 March 2022 13:26 UTC](https://fluka-forum.web.cern.ch/t/multiple-cylindrical-source-as-shown-in-figure/2806/7 "2022-03-20T13:26:26Z")