Speaker
Description
The complexity of the accelerator facility demands a high level of automation to maximize the time for physics experiments and reduce energy waste. CERN and GSI collaborate on AI model development using the commonly developed GeOFF software as part of the EURO-LABS project for real-time optimization of beam parameters and experimental setups. In the iRIS project, integration of GeOFF with the EPICS control system enables deployment in local contexts at more than 30 accelerator and telescope facilities. The eco-design driven by the FCC with more than 160 institutes and by ET withmore than 50 institutes worldwide helps to spread the methodology at global scale.
The capabilities of Geoff for energy-efficiency improvements will be demonstrated within the IRIS project at CERN, GSI/FAIR, and GANIL. Possible applications include hysteresis elimination, beam-loss reduction, and cooling optimization. The capabilities of GeOFF will be further demonstrated by optimizing the operation of the CIME cyclotron at GANIL and extending bayesian optimization to physics driven reinforcement learning.
| Article in EPJ journal as Open Access, peer-reviewed publication | No |
|---|---|
| Participation in Early Career Resarcher Award | No |
| I am an Early Career Researcher according to the ECFA definition | No |
| This research has not been previously published and is not under review elsewhere | No |
| This work is genuinely our own | Yes |