Fast Beam Tuning for Accelerator Driven Systems
The biggest challenge for Accelerator Driven Systems (ADS) is the stringent availability requirement of >99% compared to ~80% achieved by a typical accelerators. In addition to overall availability, due to thermal stress problems, ADS is also sensitive to the length of each downtime. A significant source of downtime is re-adjustment tuning of the system to account for drift in component behavior, or substitution of a backup device for one that failed. Tuning at present is done by hand, i.e. with human observation, interpretation, and decision, a process which takes hours; whereas ADS requires recovery in minutes. In this research, we apply intelligent controls in a (simulated) proton linac to automate fine-tuning. Beam monitor data is fed into a controller which adjusts magnet currents and RF power to minimize beam loss. We consider fluctuations in ion source characteristics; drift in magnet behavior (mechanical motion, or change in calibration); and failure of an accelerating cavity.