Stage de M2 : Mining Arronax accelerator data for anomaly detection

ARRONAX (Accélérateur pour la Recherche en Radiochimie et Oncologie à Nantes Atlantique) is a particle accelerator located in Saint-Herblain, Arronax produces innovative radioisotopes (radioactive atoms) for nuclear medicine research. Its main application domains are oncological diagnostic and therapeutic imaging, as well as oncology, through PET (Positron Emission Tomography). ARRONAX is a joint initiative between Subatech lab and the CRCINA (Centre Recherche en Cancérologie et en Immunologie de Nantes (CRCINA).

Abstract :

Providing precise irradiation for radioisotopes relies on the knowledge of the impact of the most effective settings of the accelerator. This is essential to secure irradiation within some time and specifications. Towards this end, the accelerator environment has been strengthened with an additional control and acquisition system, as well as a new diagnostic tools, both bringing additional data to the large data framework EPICS (Experimental Physics and Industrial Control System).


The aim of this internship is to perform data mining on the extended data collected from the accelerator, by studying the pertinence of different machine learning algorithms to monitor the accelerator settings and thus improve secure irradiation. The primary goals are the application of the algorithms to the accelerator operation, to explore:

  • significant events and detect anomalies,
  • alarm handling and prognostics.
    Secondary goals tackle tuning and optimisation of the accelerator settings for operation.

Responsables de stage : Freddy Poirier (Arronax), Diana Mateus (LS2N)

Téléchargez le sujet complet et les contacts ici :