Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
HK: Fachverband Physik der Hadronen und Kerne
HK 3: Instrumentation I
HK 3.6: Vortrag
Montag, 30. August 2021, 18:00–18:15, H2
Machine Learning based calibration of Low Gain Avalanche Detector — •Vadym Kedych1, Wilhelm Krueger1, Adrian Rost1,4, Jerzy Pietraszko2, Tetyana Galatyuk1,2, Sergey Linev2, Jan Michel3, Michael Traxler2, Michael Traeger2, and Joachim Schmidt Christian2 — 1Technische Universität Darmstadt, Germany — 2GSI GmbH, Darmstadt, Germany — 3Goethe-Universität Frankfurt, Germany — 4FAIR GmbH, Darmstadt, Germany
Linacs suffer from high power consumption for particle acceleration when high energies are desired. Because of this there is a huge interest to accelerators with idea of energy recovery. ERL allow to recirculate beam to the main linac second time with a phase shift of 180∘ which cause to deceleration of the beam and returning energy to RF cavities. The S-DALINAC at TU Darmstadt allows the possibility to operate it in an ERL mode. Optimization of the acceleration and deceleration processes are extremely important for efficiency operation S-DALINAC in ERL mode. For these purposes setup based on LGAD are being developed. LGAD is a silicon detector optimized for 4D-tracking with timing precision below 50ps thanks to internal low gain which makes it an ideal candidate for precise timing monitoring at S-DALINAC.
In this contribution we present status of a machine learning based calibration for LGAD using deep learning and neural network (NN). Experimental data from proton beam run at the COoler SYnchrotron (COSY) facility in Jülich is used to train the calibration model.
*This work has been supported by DFG under GRK 2128.