Grant: Subatomic Physics Envelope - Project

Investigations at the frontier of Nuclear Physics require rare isotope beams drive a variety of nuclear physics experiments, probing unstable and exotic atomic nuclei to high precision. Modern heavy ion accelerators are essential tools for the delivery of these beams like the linear accelerator (linac) complex ISAC at TRIUMF. With the Advanced Rare IsotopE Laboratory (ARIEL) completion, new target stations will come online. Then machine tuning at ISAC will be more challenging due to multiple beam operation and increased demand for novel beam compositions and properties. Reducing tuning times directly translates into an increased experiment luminosity on a highly subscribed and competitive schedule. A collection of high-level applications (HLAs) are being developed for use by operators and to move away from tuning the accelerator manually. Beam optics of rare isotope separators, accelerators and experiments at ISAC-I and -II and most recently the CANREB and ARIEL systems at TRIUMF have relied on sophisticated codes, tracking a population of charged particles through fields. While these illustrate individual particle evolution, they are generally computationally intensive and slow. Envelope codes such as TRANSOPTR, offer an alternative means. By tracking the reference particle and second moments of the beam, a significant computational advantage is gained, enabling real-time simulations. Beam properties are extracted on-line using diagnostic, automatically fed into the envelope simulation, which does compute the optics settings of the accelerator and beamlines. The optimized parameters are fed back into the machine, which is the basis for Model Coupled Accelerator Tuning (MCAT). An emergent application of artificial intelligence (AI) in accelerator physics is automated beam manipulation in accelerators that run high beam power. MCAT and TRANSOPTR are the ideal tools for training of the AI agent and a first test at a low beam energy transport line has been successfully completed. The proposed project will complete the thorough benchmarking of the TRANSOPTR model of the full ISAC accelerator also including the new ARIEL and CANREB systems. A salient example of the activity is the optimization of beam energy and beam spot size at the position of nuclear physics experiments. Once established at the accelerator, we plan a roll-out of advanced beam tuning to the DRAGON experiment at ISAC-I. TRIUMF's unique DRAGON (Detector of Recoils And Gammas of Nuclear Reactions) facility is the only one in the world capable of experimentally measuring many of important astrophysical reactions in stars and explosive events. The spectrometer is a complex system to tune and requires significant set-up time. The final goal of this three-year project will be the demonstration of on-line, AI driven beam tuning of the ISAC-I linac and beamlines, for example allowing for rapid and frequent changes of beam parameter to nuclear physics experiments.