Add VELO material model to RapidSim and ability to use it to compute Brems

[particleist]

We should incorporate the VELO material model from Mike&Phil as an option. Once this is done we can have Brems which properly correlates to a track's position in the VELO, and we can then smear and save the Brems photons which the material model creates as it goes along. This will also open up the ability to smear tracks according to their multiple scattering in the VELO, and thus have a proper smearing of SVs. I'll make a new branch and dump the code I got from Phil onto it, and we can then start discussing how much refactoring it will need, since the coding style and conventions have nothing to do with the ones used in RapidSim...

[bifani]

Awesome!

I remember you presented some preliminary result at an R(X) meeting that showed this method systematically generates too little brem compared to what is in the full simulation
Is this still the case?

You also mentioned two possible reasons:

1. missing material (TT/RICH1)
2. mismodelling the brem for a given material

For 1), the Birmingham group developed an MC test where one can study the amount of material traversed by a particle between two planes placed in the detector as a function of e.g. (x,y)
I still need to understand how flexible and up-to-date the tool is, but do you think this could be a useful, although simplified, parameterisation of the material after the velo?

[particleist]

Yes this still undergenerates Brems. I think the most important thing for a RICH/TT/UT model is that the format of how this is described is the same one which Phil uses for the VELO, so that it can be plugged in as "just another layer" into the code. Then how the model is generated into this data format is another story, if you have a tool which can be used for this fantastic. I would suggest you liaise with Phil and also Yasmine's student Vitalii who is looking at this I think.

[bifani]

ok, I don't think that this tool can produce a model as detailed as Phil's tool but I will check with Nigel and the students to better understand what can be done

[particleist]

I don't think it needs to be as detailed necessarily, we can approximate the TT/UT and RICH as single blocks of a certain average thickness I would have thought?

[gcowan]

This is what I was thinking. Should be good enough.

[bifani]

sorry, I probably misinterpreted what you meant by "same format as the VELO"

what I think can be done is to produce an (x,y) map containing the average material traversed by a particle between two z coordinates
depending how easy/feasible it is to move the sampling planes it might also be possible to e.g. parameterise independently TT and RICH

[particleist]

"same format" means the same root file format as the root files from Phil describing the VELO geometry, nothing more. I don't actually know what that is. :)