I have a FreeCAD GDML workbench that I would be grateful if people could try out and evaluate. It still needs further work but in its current state one can load all be it very slowly the sample file lhcbvelo.gdml which I believe originally came from CERN.
Rather than convert GDML solids to FreeCAD BREP objects FreeCADs ability to implement objects as Python features is used to implement GDML solids. So by activating the workbench one is offered a number of icons. Selecting one of the icons means a GDML solid is created. One can then change the solids attributes which are the same as in the GDML User Manual via fhe FreeCAD facility to change an objects properties. At present only a subset of GDML solids is implemented, basically all those used in the lhcvelo.gdml file. The plan is to implement more if people think the workbench might prove to be useful. I would like to get some feedback before pressing on.
There is also a sister workbench that uses the ability to call Geant4 from Python. If the plain GDML workbench proves useful then the plan is to look at running simulations under FreeCAD.
Have found a number of sample GDML files at https://github.com/JeffersonLab/remoll which have shown up a number of bugs. Will start trying to address these in a week or so’s time
Update
Seems to now be okay with
remoll / geometry_Mainz / detector_Mainz.gdml
remoll / geometry_Mainz / detector_Mainz.v2.gdml
Going to start work testing
remoll / geometry_sculpt / hybridDaughter_merged.gdml
which needs tessellated function to be added at least.
Latest version should now support loading of a default GDML file and export should create GDML definitions for constants, isotopes, elements, materials the same as imported either from the Default or selected GDML file.
Okay assuming people create a GDML file to allow Monte Carlo simulations via ROOT or Geant4. Is there any mileage in being able to output a STEP or FreeCAD file?
With the workbench implementing GDML solids as FreeCAD Python Objects the created solids therefore have a FreeCAD shape (i.e. OCC shape) for display as well as the GDML solid details. It would therefore not be hard to add the facility to export a STEP or non GDML FreeCAD file. If you think this would be useful please could I trouble you to like this post, so I can gauge response.
First of all I think this is really nice project and I am definitely following this!
I tried to import some of my handmade GDML files, but it crashes on Nist materials definition. I tried also to make few solids, export and import again, but it does not work at all.
Is there some specific thing I should test for now?
Thanks for your interest.
Are using the latest version on github?
Please could you send some of your handmade files, also exports of the few solids you have tried and do not import correctly. email is at the bottom of the readme.
In SampleFiles you should find
Subdirectory GDML_Solids of SampleFiles should load or report unsupported solid.
Subdirectory CERN has a file lhcbvelo should load although it does take sometime and you have to click on the first wait.
The ones in Alice subdirectory are very large and I have yet to manage to load.
Okay Changing to use FreeCAD Part for Volumes so that hopefully STEP export would retain structure broke a lot of things.
Hopefully they have now been fixed, so if somebody has a small GDML design that they could check if it exports to STEP okay, then that would be appreciated.
I know the main use of GDML is to allow Monte Carlo simulations but do people ever need to perform Finite Element Analysis on the same structures?
Analyses of Heat Transfer comes to mind.
I have added an experimental branch called compound to the github repository with the intention of facilitating use of the FreeCAD FEM workbench with GDML.
There is now another experimental branch called scan.
Very large files like Alice.gdml take far far to long to load. So switching to the scan branch means that the file is initially only scanned to a limit volume depth.
You can then expand individual un-expanded volumes by selecting them and clicking on the expand volume.
Would appreciate feedback on this and the compound branch.
note: There is also a branch named comp+scan which is a merge of compound and scan.
Was hoping that my icon designer would have done the business for the new
icons, but for some reason he has not come back to me so at present they are both X’s
You can find a version of the CMS geometry in the VecGeom repository along with a set of simple examples. A part of LHCb geometry is available on the CAD-DGML website.
