Dear @ted.liang ,
Congratulations on the excellent geometry models!
We face a similar challenge when modelling LHC or other CERN beam lines.
Our approach is to model the vacuum chamber in the initial FLUKA file including the tunnel and then place the magnets (or other beam elements) as replicas, as you have done.
This ensures continuity between different vacuum sections while tolerating minor mismatches in the vacuum chamber, particularly in the interconnections and in the arcs.
To illustrate this, take a look at the attached image which shows LHC Insertion Region 7 with magnets and collimator as replicas superimposed on continuous vacuum pipes.
I hope this information is helpful.
Cheers, Luigi
