Quarks and gluons cannot be observed as free particles in the detector, but form colour-neutral hadrons, which produce a particle shower in the electromagnetic and hadronic calorimeter. The resulting particles deposit energy in the individual calorimeter cells. To ensure that the cells have detected a real signal and not just noise from the detector, cells are first grouped into topoclusters. The energy of the topocluster is compared with a threshold value in order to suppress noise. At the LHC, there is not just one collision every 25 ns but up to 70 collisions simultaneously, but only one collision is of interest, which contains the hard process (e.g. ttbar production). The other collisions lead to additional energy in the calorimeter, which should be removed before or during topocluster formation if possible.

In this project, initial studies will be carried out to identify and, if possible, remove energy depositions of pileup in the calorimeter. Furthermore, studies on the formation of topoclusters can be carried out using machine learning in order to improve the algorithm.

Back