Geometry visualisation in BDSIM : errors when importing GDML of complex geometries

[Eliott Ramoisiaux]

In the context of the CERN Proton Synchrotron extraction pipe modelisation (MU16), the python library pyg4ometry was used to create the GDML for the external geometry by using tessellated solids and unions of extruded solids.

Therefore, a complete model of the extraction pipe was modelled in pyg4ometry without errors and extracted in GDML format.

Unfortunately, errors appear in BDSIM when importing the GDMLs of the different part of the extraction pipe.

A small PowerPoint presentation has been realised to present the different part of the extraction pipe model and their associated error.

BDSIM input files are also provided to allow the reproduction of the different errors.

[Laurie Nevay]

Hello,

So one issue is actually Geant4’s handling of visualising Boolean solids. The first errors where the Qt visualiser can’t display anything or that particular solid is a long standing issue in Geant4. From memory, for example, a Boolean with a primitive and a tessellated solid won’t visualise. We’re working on a fork of Geant4 to improve this - I’ll message you directly.

Another potential issue is the navigation problem. When using the ray tracer, it’s really tracking photons around, so when you see problems here, it’s really a tracking problem for real physics. Given you’re using a tessellated solid, this is likely due to it not “holding water”. If there are any holes in the solid and a particle goes through a hole, this would happen. When you’re making for example, tubes or cylinders, make sure to completely close the solid at each end. If you want a hollow cylinder, you would have to make two completely cylinders and subtract them, or write the tessellation of the end faces and inside also.

The crash on slide 7 (top right image) is interesting. Which version of Qt, XCode, clang, Geant4 are you using?

[Laurie Nevay]

Hello Eliott,

Is this issue resolve now? ie was it bad input tesselated solids + geant4 bad boolean visualisation?

Cheers,

Laurie

[Eliott Ramoisiaux]

Hello Laurie,

Yes, it is resolved and it was well both a bad input tessellated solids and geant4 bad boolean visualisation.

The correct tesselated solids creation workflow is on the pyg4ometry branch feature_tessellated.

‌

[Laurie Nevay]

Thanks Eliott. I'll close the issue now.

Cheers, Laurie