Skip to main content


  • 12 December, 2018

A new era of predictive maintenance in railway is dawning

Every day millions of people travel by rail, whether for commuting, business, or pleasure. Thousands of people are also inconvenienced by delays and cancellations resulting from the failure of one or more rail assets or engineering works.

Train operators and infrastructure owners strive to provide the best service they can while maximising safety and profit for all involved. As cancellations and delays incur penalties and reputational damage, huge amounts of money are spent every year on scheduled preventive maintenance – maintenance that is performed through a cyclic set of actuations, ensuring the perfect state of the whole railway network infrastructure, at all times. However, these actuations may sometimes be more frequent than needed, incurring higher maintenance costs than necessary. What’s more, preventive maintenance doesn’t allow the operator to accurately assess the probability of a railway infrastructure failure.

Better rail maintenance envisioned

The rail industry has invested significantly in research to improve rail trackside asset health and related maintenance planning strategies. The TransformingTransport (TT) research project, partially funded by the European Commission, aims to radically improve maintenance planning of critical rail assets.  The objective of the Proactive Rail Infrastructures work package within the TT project is to revolutionise how faults and wear of rail assets are identified, providing rail infrastructure managers and operators with new predictive maintenance systems.

This is needed to reduce operational costs, minimise traveller service delays, and deliver workforce safety improvements, supported by data collected from the infrastructure and from existing knowledge, then customised for each rail asset.

Over the past 18 months, Thales UK has been leading the proactive rail infrastructures work package within the TransformingTransport research project, which consists of two pilots; the first is in the UK focussing on a mainline rail line, and the other in Spain focusing on a high-speed line in Cordoba - Malaga.

UK pilot projects helps pinpoint rail faults more efficiently

The UK pilot is referred to as the initial pilot and covers three areas (or use cases): point machine/track circuits, overhead line equipment, and train-to-track interface. A team of data scientists from Thales Research Technology and Innovation, the Thales R&D lab CeNTAI (Centre de Traitement et d’analyse de l’Information) and the University of Southampton have been running Big Data analytics across vast quantities of historical data sets in these use cases. 

Valuable insights into the causes of rail asset failure have emerged, rapidly improving diagnosis and prognosis, in particular regarding the point machine where it is now possible to identify different types of fault characteristics for an asset.  The results of the fault classifier algorithm were presented to the European Commission during a live demonstration in Valencia, Spain, in September 2018. Network Rail, the UK Infrastructure owner and end user of this pilot, recognised the value of the outputs, considering them an important milestone in the development of the system.

Spanish pilot project advances rail maintenance

The Spanish pilot also consists of three use cases: point machine, track profile, and operational restrictions. The pilot is improving the reliability of high-speed rail networks by optimising operator performance and maintenance of the rail infrastructure. It uses Big Data technologies to understand the variables that impact operator performance and to model maintenance activities performed in the infrastructure (tracks, tunnels, bridges etc.). This is based on data from rail traffic, rolling-stock flows, maintenance logs, planning and control activities and other information sources.

Similar to the UK pilot, the Spanish main line rail infrastructure companies are working together to improve cost efficiency and capacity through

The analysis will allow the operator to anticipate the maintenance activities on the rail network and improve maintenance operations across the whole rail infrastructure. It will enable rail operators to predict in real-time the impact of certain rail-related occurrences and events on traffic management.

Both pilots have ultimately created Big Data solutions that could bring radical improvements in rail maintenance, leading not only to significant savings but also to more efficient transport for both operators and passengers.

Our Tought

Add Your Review