Modern train and communication network concept

Turning a passenger train into an infrastructure monitoring train: a Trial Infrastructure & Algorithm Development (KEEPSAFE 2)


Network Rail

Value to Coventry University


Project team

Professor Alexeis Garcia-Perez


Coventry University (lead), Network Rail, Virgin Trains, Alston Transport, Serco

Duration of project

01/02/2014 - 31/12/2014

Project overview

Building on the outcomes of the project ‘A KnowledgE Elicitation aPproach to understanding railway SAFEty’ (KEEPSAFE 1), this project aimed at creating a data-driven system-view of the railway overhead line infrastructure. The research was a live trial of the rail industry’s desire to use sensors fitted to passenger trains to measure the system interactions between train and infrastructure. 

The consortium worked to develop a prototype solution that served as a demonstrator of the feasibility of turning every train into an infrastructure monitoring train. In doing so, the outcomes supported a greater understanding of those interactions at a system level and a move from a find and fix culture to predict and prevent intelligence across the railway network.

Project objectives

1. Eliciting and modelling knowledge from the team on health and usage of overhead line equipment (OLE) and related thresholds.

2. Collecting data from a passenger train and transferring it to Coventry University for analysis.

3. Hosting of data on Coventry University infrastructure, ensuring a secure and sustainable archiving and data management process.

4. Development of a prototype system that supports a predictive maintenance strategy for the OLE infrastructure throughout the London North West line, based on:

  • The development of predictive algorithms that validate the data, as well as provide statistical confidence in the point measurement of key OLE parameters (force and acceleration) at a specified location.
  • The implementation of a software prototype that visualises the data across the London North West line so that Network Rail engineers can take appropriate action in terms of predictive maintenance.

5. Delivery of source code and database for integration in the Network Rail On-Line Analysis & Processing System.

  • Using the data collected from a passenger train a prototype solution for rail infrastructure monitoring and management was developed by the Coventry University team and its validity demonstrated in an operational environment (to technology readiness level 6).

    Not only were actual risks identified, but the approach could be rolled out across the rail network, pre-empting system and network failures, reducing service interruptions and mitigating operating risks. This led to a digitally-enabled enhanced infrastructure management strategy for greater efficiency, safety and reduced maintenance costs.

  • Garcia-Perez, A., Shaikh, S.A., Kalutarage, H.K. and Jahantab, M. (2015) ‘Towards a knowledge-based approach for effective decision-making in railway safety’. Journal of Knowledge Management 19 (3), pp.641-659. DOI:

 Queen’s Award for Enterprise Logo
University of the year shortlisted
QS Five Star Rating 2023