ATLAS Detector Monitoring with Jets
Not peer reviewed
MetadataShow full item record
The purpose of this thesis was to provide a high level software framework for detecting large problems with the ATLAS detector at the LHC and estimate the jet resolution from the f irst data. The result was the program DiJet, written in C++ by Kent Olav Skjei and Thomas Burgess. DiJet can be run both as a program within Athena (ATLAS software framework) or as a script in AthenaROOTAccess. We have looked at and compared several variables for back-to-back jets using data sets with Monte Carlo dijets, Monte Carlo top and the simulated FDR2 data. Dijets were chosen due to the high cross section expected at the LHC. It became clear that for real data, one will need to review the dijet selection criteria in order to supress multijet background. DiJet also reconstruct the W mass as a method for estimating hadron resolution. This method requires more data than methods for back-to-back jets do, and was not useful for FDR2 data. However, based on Monte Carlo dijet and top events we were able to compare jet resolution estimates for the methods for back-to-back jets and the method based on the W mass. As expected, these variables seem to be related. In connection with the back-to-back jets we saw a certain \eta dependence of transverse energy. The last variables we looked at were missing transverse energy and the vectorial sum over jet momenta. From the latter results, it became apparent that missing transverse energy contains more than just the sum over jet momenta. And it also became clear that the jet reconstruction algorithms themselves create some false missing transverse energy. The position of bad eta and phi regions were found by looking at profile plots of these variables. Based on this we decided to only look at the region -3\leq |\eta | \leq 3 due to low statistics in high |\eta |. We have seen that the methods for estimating jet resolution in this thesis gives half the value of the mass resolution of particles decaying into jets. Based on the results, we seem to have found an asymmetry in \phi in the results from the jet reconstruction. We did not have the opportunity to investigate the cause of this asymmetry, due to lack of time. The original objective of this thesis was to look at the Express Stream. This turned out not to be possible due to the decision of not making the Express Stream generally available. In addition, this thesis has been affected by the delay of the LHC schedule, changes in the policy concerning streaming and evolution of the ATLAS software frameworks. Upgrades on the local computer cluster used for data analysis have also caused considerable loss of time.
PublisherThe University of Bergen
Copyright the author. All rights reserved