implement ElasticModel

To debug tracking and investigate UHE behaviour it could be useful to have a purely elastic model, i.e. a simplified interaction with fixed cross section, that just changes the direction of the particle. For protons this leads to a random walk through the volume.

We may have a physical dsigma_el/dt \propto exp(-Bel|t|) cross section at least as an option.

assigned to @mreininghaus

created branch 107-implement-elasticmodel

mentioned in merge request !59 (merged)

I see, you started a branch here. Please note that this Issues is within lowest in priority of all of them... This will be nice to have, but I don't think it will bring any real benefit right now.

I think it is very useful to have it early because a randomly walking particle might trigger all kinds of effects in the tracking which we didn't think of yet. Think of it as some kind of Fuzzing.

Let's see. This does not convince me... we have a lot of physical fuzzing already there. But ok: if we implement ElasticModel, it should in my opinion be a physical model, too. Elastic scattering for hadrons is something theoretically very well understood. If we do not want to have a physical model, it should also NOT be called ElasticModel. We could have a RandomWalk model or something like this for this purpose. Or a DiffusionModel, ... but would those be "Tracking" mdodules or "Processes"?

Here is the result of a first prototype. There are still a few strange issues to solve, though...

This is really amazing. I am wondering to what level of magnitude of turbulent magnetic field this would be correspond. Nobody has ever considered that for cosmic propagation calculations...

Are you sure? I guess if somebody uses sth. like CRPropa with SIBYLL, SIBYLL will also perform elastic interactions from time to time, won't it?

No, all models are inelastic event generators.

mentioned in commit aae0786d

closed via merge request !59 (merged)