PhD Studentship: Optimizing Railway Maintenance: Minimising network disruption

Making effective railway maintenance plans requires a near real-time understanding of the health of assets.  When assets degrade this increases the risk to network availability and safety.  

Over recent years there has been a surge in the use of built-in, train-borne and air-borne sensors capturing data and imagery.  These sources provide different views of the current “state” of an asset.  The availability of such diverse data provides an exciting opportunity to explore how cross-fertilisation of sources could drive the next generation of intelligent maintenance systems for the rail industry.

The track system – that includes earthworks, drainage, S&C and the vegetation that surrounds this – is a key driver:  Track maintenance cost alone is 50% of the total cost of maintenance. In the last 2 years, 50% of asset-failure delay costs were attributable to track. Consequently how the track is managed is both a key driver of cost and network disruption. This project focusses on the exploitation of sensor data relevant to the track system. 

The goal would be a demonstrator that exploits a combination of data monitored at different frequencies processed using the latest generation of data-mining and deep learning methods based on powerful GPU systems. Data sources include:

  • Track geometry monitored several times per day from in-service trains where available.
  • Environmental data monitored several times per day from external systems
  • Track geometry and rolling contact fatigue measured periodically by NR fleet
  • Condition of earthworks, drainage and vegetation based on infrequent surveys/lidar imagery.
  • Historic delay and failure incidents.
  • Network traffic/infrastructure utilisation 

 

Monitored data is now available to measure the actual condition and rate of change of condition as a consequence of actual track use.  Advanced analytical techniques can be applied to the available datasets to search for predictive models of how condition will change with use.

The project would measure the benefits in moving from a static plan for scheduled maintenance to a dynamic risk-based plan that is continually refreshed based on measured and predictively modelled changes in the track system.

This project will be undertaken in conjunction with our Centre for Doctoral Training in Next Generation Computational 

Modelling and Network Rail, who will provide for an extremely attractive and fully competitive tax-free stipend.

 

Candidates for this exciting role would:

  • Demonstrate an enquiring mind with a relentless drive to seek new insights from data
  • Demonstrate knowledge of statistics and computational techniques
  • Want to develop skills in data science
  • Want to develop skills in information technology to effectively exploit big data
  • Want to contribute to the Digital Railway vision.

 

If you wish to discuss any details of the project informally, please contact Prof Andy Keane, CED research group, Email: ajk@soton.ac.uk, Tel: +44 (0) 2380 59 2944.

This project is run through participation in the EPSRC Centre for Doctoral Training in Next Generation Computational Modelling (http://ngcm.soton.ac.uk). For details of our 4 Year PhD programme, please see http://www.findaphd.com/search/PhDDetails.aspx?CAID=331&LID=2652 

For a details of available projects click here http://www.ngcm.soton.ac.uk/projects/index.html