Mean beamstrahlung model inputs

[xbuffat]

The 'mean' model for beamstarhlung radiation relies on different inputs that the 'quantum' model. It is easy for the user to create an element for which the two models are not equivalent by setting non-consistent inputs. One should instead compute both internally based on the same input.

[xbuffat]

Mean BS model is actually very bad, it neglects the crossing angle. Probably one could find a better approach by using Yokoya formula on the full interaction rather than slice-by-slice neglecting the offset between the slices.