Discard splash events with nVeloCluster GEC

PVz peak @ -100mm was observed with PbPb + SMOG, due to the so-called (already in Run2) splash events. These have very high occupancy in the VELO and cannot be reconstructed/used for physics.

A nVeloCluster cut at 15k (safe wrt central PbPb collisions, see the plot below) is added with MR to prefilter physics trigger lines. GEC was tested against simulation (30-90% centrality) and no difference in trigger lines is observed (see enclosed logs)

output_Allen_nVeloClusterGEC.txt output_Allen_master.txt diff.txt

cc @cagapopo