Speaker
Description
The Isotope Separator and ACcelerator (ISAC) facility at TRIUMF supplies both stable and rare isotope beams for a variety of nuclear astrophysics experiments. One of these, the Detector of Recoils And Gammas Of Nuclear reactions (DRAGON), investigates reaction rates of astrophysical processes via radiative capture measurements. Currently, rare isotope beams delivered to DRAGON are manually tuned by operators—a process that is both time consuming and difficult to train for, especially given the boundary condition of high demand for beam time. This work presents a semi-automated approach to optimize beam transport through the ISAC-I linac and towards DRAGON. The method decouples the tuning of quadrupole lenses and corrective steerers. Quadrupoles are adjusted using Model Coupled Accelerator Tuning (MCAT) to match a design tune, while Bayesian Optimization for Ion Steering (BOIS) is used to do the beam orbit correction. BOIS treats steering as a black-box optimization problem, evaluating functional values only through direct measurement to maximize beam transmission. By combining MCAT and BOIS, this method offers a more efficient and physics grounded tuning process for the facility.
| Email address | ohassan@triumf.ca |
|---|---|
| Supervisor's Name | Oliver Kester |
| Supervisor's email | okester@triumf.ca |
| Funding Agency | NSERC |
| Classification | Machine Learning and AI |
