The University of Liverpool, Liverpool, United Kingdom
STFC/DL/ASTeC, Daresbury, Warrington, Cheshire, United Kingdom
Cockcroft Institute, Warrington, Cheshire, United Kingdom
Successful user experiments at particle beam facilities are dependent upon the awareness of beam characteristics at the interaction point. Often, properties are measured beforehand for fixed operation modes; users then rely on the long-term stability of the beam. Otherwise, diagnostics must be integrated into a user experiment, costing resources and limiting space in the user area. This contribution proposes the application of machine learning to develop a suite of virtual diagnostic systems. Virtual diagnostics take data at easy to access locations, and infer beam properties at locations where a measurement has not been taken, and often cannot be taken. Here the focus is the user area at the planned Full Energy Beam Exploitation (FEBE) upgrade to the CLARA facility (UK). Presented is a simulation-based proof-of-concept for a variety of virtual diagnostics. Transverse and longitudinal properties are measured upstream of the user area, coupled with the beam optics parameters leading to the user area, and input into a neural network, to predict the same parameters within the user area. Potential instrumentation for FEBE CLARA virtual diagnostics will also be discussed.
