Re: Problem with REC body

From: Joseph Comfort <Joseph.Comfort_at_asu.edu>
Date: Fri, 06 May 2011 15:03:45 -0700

I have some more information with regards to the REC bodies.

I did a side calculation and expanded the Height and 2nd Semiaxis values
to 14 significant digits (with IDBG=0 still). The full geometry of my
case was then handled (including all of my ARBs, etc.) without problems,
and the case ran without glitches. Only the REC cards needed to be
modified.

The pipe for the second pair of RECs in the test file is actually
circular, so I replaced the two REC bodies with RCC cylinders. There
were no significant changes (beyond the random number sequencing) in the
results. (Good.)

However, the final results have notable differences from those with the
2008.3 Fluka. I save beam profile data at various places along the
beamline structure. There are no significant differences between the
two Fluka versions at locations prior to the start of the Pipe1 vacuum
pipe.

But I begin to see real differences even before the end of Pipe1. (The
collimated neutral beamline is entirely outside of Pipe1, but halos
extend inside it.) At the end of the entire beamline in the
calculations, the new Fluka has about 8% less yield than the old one.
The yield within a circle of radius 5 cm is about twice as much less.

It is not obvious to me which code version is giving the most reliable
results. Was the REC coding modified? The extra sensitivity on
perpendicularity suggests it was. But, then, was a bug introduced or
was a bug fixed?

Thank you,
Joe Comfort
Received on Sat May 07 2011 - 13:34:20 CEST

This archive was generated by hypermail 2.2.0 : Sat May 07 2011 - 13:34:29 CEST