Neural Networks for anomaly detection in CO2 cooling systems

by Michal Zbigniew Zimny (CERN)

Wednesday 5 Jun 2024, 11:00 → 12:00 Europe/Zurich

32/1-A24 (CERN)

During the long shut-down planned in the years 2026-2028 (LS3) the LHC accelerator and its experiments will undergo an important upgrade phase (referred to as “phase-2 upgrade”). One of the important upgrades foreseen is the complete substitution of the tracking detectors of ATLAS and CMS and the End-Cap Calorimeter of CMS for new silicon-based technology together with a new HGTD detector of the ATLAS experiment. For the new silicon detectors of the experiments the total power dissipation is expected to be several hundred kilowatts and the detectors must be operated at a temperature below -30 ^oC. For the thermal management of these detectors, a CO₂ cooling system based on parallel operation of several modular units is foreseen.

In this seminar, a proposal for deployment of anomaly detection for the planned CO₂ cooling systems will be presented. Participants will be introduced to basic concepts of neural networks and their applicability to anomaly detection in industrial systems. 

Anomaly_Detection_Using_Neural_Networks_062024_ppp.pdf

Anomaly_Detection_Using_Neural_Networks_062024_ppp.pptx