Cargando…

Feasibility of Hardware Acceleration in the LHC Orbit Feedback Controller

Orbit correction in accelerators typically make use of a linear model of the machine, called the Response Matrix (RM), that relates local beam deflections to position changes. The RM is used to obtain a Pseudo-Inverse (PI), which is used in a feedback configuration, where positional errors from the...

Descripción completa

Detalles Bibliográficos
Autores principales: Grech, Leander, Alves, Diogo, Jackson, Stephen, Valentino, Gianluca, Wenninger, Jorg
Lenguaje:eng
Publicado: 2019
Materias:
Acceso en línea:https://dx.doi.org/10.18429/JACoW-ICALEPCS2019-MOPHA151
http://cds.cern.ch/record/2772240
Descripción
Sumario:Orbit correction in accelerators typically make use of a linear model of the machine, called the Response Matrix (RM), that relates local beam deflections to position changes. The RM is used to obtain a Pseudo-Inverse (PI), which is used in a feedback configuration, where positional errors from the reference orbit as measured by Beam Position Monitors (BPMs) are used to calculate the required change in the current flowing through the Closed Orbit Dipoles (CODs). The calculation of the PIs from the RMs is a crucial part in the LHC’s Orbit Feedback Controller (OFC), however in the present implementation of the OFC this calculation is omitted as it takes too much time to calculate and thus is unsuitable in a real-time system. As a temporary solution the LHC operators pre-calculate the new PIs outside the OFC, and then manually upload them to the OFC in advance. In this paper we aim to find a solution to this computational bottleneck through hardware acceleration in order to act automatically and as quickly as possible to COD and/or BPM failures by re-calculating the PIs within the OFC. These results will eventually be used in the renovation of the OFC for the LHC’s Run 3.