Re: [fluka-discuss]: nested geometry directives

From: Vittorio Boccone <dr.vittorio.boccone_at_ieee.org>
Date: Sat, 20 Sep 2014 18:15:33 +0200

Dear Tim,
 the "Start_translat" is optimized for speed as il easily can move a body
in space without changing its own definition, as it would happen for
example if you would rotate a RPP.

A start_transform is actually what you need.

You can define then a multicard ROT-DEFI and split there all the logical
step to move your body in the order in which you want to be executed (you
can try it, split it and inverted it easily with flair/geoviewer). What you
need at the hand is just a rotation matrix which is easily handled by
different programs.

For example. For one of the project I worked on in the past I had to deal
with complex geometries that were often changing. The ROT-DEFI were
included in an external file and we had script which was recalculating the
position of the objects and rewriting new ROT-DEFIs when the position of
the object was changing.

Best
Vittorio


On Fri, Sep 19, 2014 at 4:48 PM, Timotheus Cooijmans <tim.cooijmans_at_cern.ch>
wrote:

> Hello Fluka experts,
>
> I am trying to understand the behavior of geometry directives when
> nested. At first I assumed that nesting directives would simply compose
> the corresponding transformations, in other words the directives would be
> applied from innermost to outermost. However, I now see that the manual
> says:
> [...]
> I guess the solution would be to apply an inverse rototranslation like so:
>
> $Start_transform -<translation by -dx -dy -dz and then some rotation
> around (0, 0, 0) which is the inverse of the rotation I actually want to do>
> $Start_translat -dx -dy -dz
> ...body definitions...
> $End_translat
> $End_transform
>
> but it's confusing, and this is only one particular class of composite
> transformation I'm trying to implement; I was hoping to be able to generate
> geometry directives programmatically without having to worry about
> contextual complications.
>
> Do I understand the situation correctly? Is this the kind of thing I have
> to do?
>
> Many thanks,
> Tim
>
Received on Sat Sep 20 2014 - 19:54:35 CEST

This archive was generated by hypermail 2.3.0 : Sat Sep 20 2014 - 19:54:36 CEST