PNNL

Abstract

We describe a method for precisely regulating the Gradient Magnet Power Supply (GMPS) at the Fermilab Booster accelerator complex. We demonstrate preliminary results based on a surrogate model trained from real accelerator data. We additionally show how the neural networks that will be deployed for control purposes may be compiled to run in an on-board environment using field-programmable gate arrays (FPGAs). This capability is important for operational stability and working in real-time environments like an accelerator facility.