Artificial Intelligence for Predictive Maintenance - Executive Report Year 1 (2022)

One of the major challenges of the Railway System is to maximize system availability, to increase the level of information and to continuously improve the safety level of its mobility services. The maintenance operations for Infrastructure, Rolling Stock and Signaling strongly contribute to these objectives. Up to now the main practices to manage maintenance were breakdown-based, time-based or condition-based preventive maintenance. Predictive Maintenance will be the next level of maintenance practices evolution.

The European Strategy for the Rail Sector “Challenge 2050” as well as the Rail Technical Strategy Europe (RTSE) both heavily promote the use of advanced technology in order to foster innovation, especially with the use of the new available techniques of Artificial Intelligence (AI). This has set a framework to encourage reflection on the use of the new technology based on AI to enhance Business in the Railway sector.

In line with this perspective within the railway domain, one of the most promising areas that needs to take advantage of Artificial Intelligence is the area of Predictive Maintenance. Knowing that the maintenance costs of the whole railway system for Infrastructure Managers and Railway Undertakings count for a major part of their recurring spending, one can anticipate that the introduction of AI could result in business benefits as described in chapter 2.2.