Using AI to speed up nuclear power plant inspections 

AI start-up ALEIA and Omexom NDT Engineering & Services, together with the Laboratoire d’Acoustique de l’Université du Mans (LAUM) UMR CNRS, an acoustics laboratory, have announced the rollout of the AUTEND project, which is intended to speed up nuclear power plant inspections thanks to artificial intelligence.

Omexom NDT specialises in the design, qualification and implementation of automated processes in non-destructive examination (NDE) and has been operating in the field of critical equipment in French nuclear power stations for over 40 years. 

With the rate and number of nuclear facility inspections rapidly increasing, the AUTEND project aims to facilitate and accelerate the work of analysts in the field through AI-driven automatic identification of areas to be inspected. 

The project focuses on non-destructive examination, in other words inspection processes for nuclear infrastructure using eddy current and ultrasonic methods. 

Applying AI to inspections helps increase analytical capacity while maintaining the reliability of result interpretation. Overall, the detection of these areas reduces analysis time. In time, AI will significantly boost the theoretical reliability of examinations, in particular through the gradual development of a scalable database. 

The project is built on the ALEIA platform and populated with appropriate (in terms of quality and quantity), anonymised test sets. Hosting is provided by a sovereign cloud, thus ensuring full control of information processing by users.  

The two-and-a-half-year project (lasting from 2022 until 2024) is supported by the French Ministry of the Economy, Finance and the Recovery and by Bpifrance, a public investment bank, as part of the Recovery Plan.