Comment

Digital railway maintenance

Tim Flower, head of maintenance at Network Rail, explains how the infrastructure owner is adopting a new approach for condition monitoring to maximise return on investment.

Over recent years Network Rail has made significant progress in delivering efficient and effective maintenance, providing the foundation for our forward strategy. For instance, substantial improvements have been made in terms of understanding and monitoring the asset base (61% of points and 26% of track circuits are now monitored, with installation based on asset criticality), resulting in service affecting failures being at an all-time low. 

Aspects of Network Rail’s infrastructure monitoring capability are also acknowledged as best in class, with the number of broken rails lower than the five-year control limit for 12 of the previous 13 periods. Over the next year, we also expect to conclude roll-out of plain line pattern recognition over approximately 15,000 track miles and introduce automated inspection of surface cracks using eddy current on 18,000 track miles. And Failure Modes Effects and Critical Analysis (FMECA) is now at the core of the process used to define and implement reliability-centred maintenance regimes. 

However, many challenges remain, including the fact that maintenance expenditure is not reducing at the rate forecast at the start of the control period. 

Additionally, service affecting failure delay is increasing despite the number of failures reducing; track access required for maintenance continues to reduce; and condition monitoring data is not yet utilised to generate predictive and preventative maintenance regimes which, once implemented, will reduce whole-life cost whilst improving asset performance.  

Effective and high-quality maintenance task delivery is a priority to maintain the sustainability and performance of the asset base. But  our datasets do not have the quality and granularity required to fully deploy Predict & Prevent maintenance. Reliability-centred maintenance regimes are also not fully optimised based on performance risk to enable trade-offs between maintenance spend and the impact of asset failure. 

A new team 

To deliver the opportunities identified, a new team focusing on monitoring (Predict & Prevent) has been introduced into the head of maintenance team within the Safety, Technical & Engineering Directorate. This brings together the condition monitoring engineering and business improvement resource as well as creating the focal point for a wider virtual team. 

To deliver the required transformation, Network Rail is utilising the following approach to prioritise the condition monitoring workbank, maximising return on investment today whilst recognising the need to address future business need: 

National cost/performance data 

There are three cost drivers Network Rail is focusing on for condition monitoring to maximise return on investment: renewal cost, maintenance cost, and cost of Schedule 8 due to asset failure. 

FMECA 

For a number of years, Network Rail has been utilising FMECA to develop reliability-centred maintenance regimes across the majority of asset disciplines. However the approach and the outputs of the analyses have not been fully embedded across our asset management system, our fault management system (FMS) and our product acceptance process. The current state is:  

• 24% of asset types currently have FMECA loaded into FMS
• 59% of key signalling assets have FMECA completed
• Approximately 70% of assets have some form of FMECA in place 

A full gap analysis of the current position is currently being performed with a target of having 80% of FMECA utilised consistently across all systems.  

The product acceptance (PA) process has been reviewed and, from April, all products requiring PA must provide evidence of meeting the new Design For Reliability (DFR) standard. The evidence to be provided includes Fault Code Lookup data tables in standard format derived from their FMECA, which enables immediate loading into FMS. The PA process also includes provision of a suggested maintenance schedule using reliability-centred maintenance based on the FMECA.

Condition monitoring technology 

Utilising the cost data and understanding the FMECA allows Network Rail to assess the priority of roll-out of future condition monitoring technologies. As stated in the introduction, both points and track circuits have mature solutions in place, alongside power supply monitoring and points heating. Network Rail’s linear asset recording capability is also very mature, with a large number of sensors fitted to our monitoring fleet. 

Network Rail has the following condition monitoring technology requirements: 

• Level crossing monitoring
• Train borne monitoring of signalling equipment
• Train borne monitoring of S&C, transferring sensors that are business as usual in plain line to automate inspection, measurement and testing of S&C. It is anticipated that up to six vehicles will be required if the technology can be proven
• Overcoming constraints with current data collection technologies (e.g. rail flaw, eddy current) to enable higher speed recording and/or fitment to in service trains.
• Use of sensors on service trains to improve asset management decision-making 

Network Rail’s challenge statements, published on our website in December, acknowledge that there will be ideas, prototypes, demonstrators or systems that are in use and/or development in other railways or industries that could provide significant benefit to the UK railway, and we would like to work with the supply chain to develop these.  

Data analytics 

Utilising data analytics as part of predictive and preventative maintenance is seen as an area with huge potential. Pilots for diagnosis of points failures and earlier prediction of track circuit failures have been developed and will be tested in the near future, but there are many other opportunities: 

• Deliver full prognostics tools for both track circuits and points
• Develop prognostic capability for all assets fitted with condition monitoring
• Development of algorithms to establish degradation rates of our linear assets
• Provision of additional insight through integration of disparate data sources. Examples identified are:

1. Integrating points condition monitoring and train borne monitoring data to understand the effect of poor track conditions on the life and reliability of points operating equipment
2. Performance risk of overhead line heights and staggers position moving away from the original design whilst remaining compliant to existing standards
3. Understanding how changes in track stiffness can be used alongside age and condition data to predict and prevent broken rails 

Network Rail is interested in understanding how analytics such as machine learning or expert systems are used in other railways or industries. Again, this is detailed in the Predict & Prevent challenge statement. 

Maintenance workbank scheduling 

The final part of the value chain is the ability to utilise the FMECA, condition monitoring and data analytics to manage the workbank. This will require the creation or configuration and integration of systems to bring together the analysis to enable systematic and regular reassessment and optimisation of maintenance regimes, to maintain asset performance to specified levels. A systematic and robust risk-assessment approach utilising cross-discipline FMECA will need to underpin these systems.

The benefits 

It is estimated that Network Rail has implemented monitoring systems and processes to approximately 33% of maintenance by cost, and it is anticipated that a maximum of 66% of maintenance can be optimised (due to asset age and condition). 

This opportunity is evidenced through studies performed by asset management consultancies, independent research by leading universities and an independent audit of London Underground engineering and asset management capability. 

These identified that reaching world-class predictive, risk-based maintenance strategies delivers the following benefits: 25-35% reduction in maintenance costs; 70%+ reduction in the number of service failures; 35-45% reduction in down time following failure; 20%+ increase in workforce productivity; and, most importantly, the above all result in fewer unplanned, reactive interventions, less ‘boots on ballast’, delivering enhanced workforce safety.

FOR MORE INFORMATION

W: www.networkrail.co.uk

Tell us what you think – have your say below or email [email protected]