Luminosity calculation framework: a summer student project at CERN

Roman Lavicka

When I recieved an email from CERN HR department in the beginning of April starting with the word "Congratulations", I was the happiest person in the city at that moment. My name is Roman Lavicka and I'm an undergraduate student in experimental particle physics at Czech Technical University in Prague. Next year I'm going to work on my master thesis under the supervision of Guillermo Contreras.

Participation in the CERN Summer Student Programme is a fantastic opportunity for every student  with a technical specialization. Here, you can work side-by-side with the best experts in the field, acquire new contacts with physicists not only from the collaboration, gain new experiences and meet new friends from all around the world. That is what I call a productive summer.

My project is to develop a new luminosity calculation framework under the supervision of Evgeny Kryshen. A motivation for this effort is to provide a unified and precise luminosity source for ALICE analyses, which can be easily accessed in various formats. This needs a sort of automatic program, which will be periodically launched. A unix-based software utility Cron is used to manage it. Also run/trigger coordinators need up-to-date information. The goal is to provide them with actual summaries on collected statistics for different LHC periods.

Luminosity calculation is performed by multiplying delivered luminosity by the lifetime of each trigger class. The delivered luminosity is calculated as a ratio of the number of counts and corresponding cross sections for a reference trigger class usually based on trigger inputs from T0. This ratio is further corrected for pile-up probability.  The trigger class lifetime is computed as a ratio of trigger class counts after and before various CTP vetoes. In order to improve precision of the lifetime calculation, we use aliases pointing to a class with the most precise trigger counts in the trigger cluster.

As a source of data we used the OCDB files. A list of runs is periodically updated and if new runs are detected, they are analyzed by special scripts. For the luminosity calculation we use only runs with PHYSICS_1 partition and stable beams. Once the run has passed these criteria, results are stored in root files and a graphical output is produced. All information is stored online on my personal webpage for now ( In the future we would like to provide luminosity and collected statistics class-by-class and run-by-run in OADB and Monalisa.




Alice Matters