Eiffage tests AI-based predictive maintenance for rail infrastructure on BPL-HSL

Eiffage Énergie Systèmes said that it is trialling an artificial intelligence (AI) system to support predictive maintenance on the French Brittany–Pays de la Loire high-speed line (BPL-HSL).

The trial is focused on identifying potential faults before they lead to service disruption.

According to the company, the initiative aims to maintain passenger comfort, ensure train safety, and extend track service life by using data to prioritise interventions.

The trial covers the 182km BPL-HSL, which opened in 2017, and links Paris and Rennes in 85 minutes.

Ferlioz, the Eiffage Concessions brand responsible for operating and maintaining the line, manages infrastructure that carries about 30,000 trains a year and reports a punctuality rate of 99.2%.

The predictive maintenance system integrates data from multiple sources.

It analyses sensor readings from SNCF Réseau’s IRIS trains, maintenance records from Ferlioz’s computerised maintenance management system (CMMS), and geometric track design and traffic data such as train counts, speeds, and tonnages.

Eiffage’s AI Expertise Centre applies a hybrid approach that combines machine learning and deep learning to detect anomalies that could indicate forthcoming failures and to inform maintenance scheduling.

Eiffage digital expertise activities manager Jean-Louis Haller said: “Managers carry out conditional maintenance on infrastructure by regularly checking track geometry parameters, such as deformation, lateral alignment, cross levelling, and rail spacing.

“However, given the challenges of availability, cost and safety, predictive maintenance offers an innovative approach using AI to anticipate failures and optimise operations.”

Concurrently, engineers are measuring the mechanical behaviour of track structures using more than 100 sensors installed across four representative sections of the line.

The instrumentation records temperature, humidity, deformation, and acceleration to supply contextual inputs for the AI models.

Project teams stated that combining algorithmic analysis with field expertise enables earlier detection of issues, improved allocation of maintenance activity, and continued adherence to safety requirements.

They also indicate that the method could be applied to other rail corridors and to different categories of infrastructure as demand for sustainable mobility grows.

In 2022, a 50:50 Colas Rail–Eiffage Énergie Systèmes joint venture won a €26m Société du Grand Paris contract to supply traction equipment for Express Line 18.