Cargando…

MD3343 - Fully-Automatic Parallel Collimation Alignment using Machine Learning

The collimation system protects the LHC and its sensitive equipment form unavoidable beam losses. Collimators are set up in the form of a hierarchy which is established using a very precise alignment of all the collimators with respect to the beam. In the past years collimators were aligned using a...

Descripción completa

Detalles Bibliográficos
Autores principales: Azzopardi, Gabriella, Valentino, Gianluca, Salvachua Ferrando, Belen Maria
Lenguaje:eng
Publicado: 2018
Materias:
Acceso en línea:http://cds.cern.ch/record/2647213
Descripción
Sumario:The collimation system protects the LHC and its sensitive equipment form unavoidable beam losses. Collimators are set up in the form of a hierarchy which is established using a very precise alignment of all the collimators with respect to the beam. In the past years collimators were aligned using a semi-automatic approach whereby collimation experts manually control the alignment of both beams in parallel. During commissioning 2018, the first version of the fully-automatic software was tested. The first version of this software did not account for crosstalk between the beams, therefore the two beams had to be aligned sequentially. After commissioning, a crosstalk analysis model was developed and integrated into the fully-automatic software. The new parallel software was tested during this MD and managed to successfully align both beams in parallel at injection for the first time.