Clever Thinking in Predictive Analytics

    Using predictive analytics technology to the best effect, Southern is taking an innovative approach towards its fleet management. With train services for commuters, airport users, business travellers, and leisure passengers, around 447,000 journeys are made every day with Southern.

    The company operates services in south London, and between central London and the south coast, through East and West Sussex, Surrey, and parts of Kent and Hampshire. Southern, in various guises has operated franchises on its network since 2001 and has invested heavily in its trains and stations, train depot upgrades, and overall performance. It has also created an environment where innovation can flourish in order to drive improvements or add value to the business.One example of where this is taking place is within Southern’s fleet directorate. At present the company is undertaking a complete refresh of its 182 Class 377 trains (700 carriages) in-house at its Selhurst Depot.

    Predictive analytics


    Given the extent of the refresh programme and tight turnarounds, clever thinking has had to be employed as fleet overhaul and projects manager, Iain Nairne highlights: “There’s not an area of the train inside or out that we don’t touch in some way or another from new flooring and seat covers, to the overhaul of the bogies and doors, and paint repairs. Work on each unit begins on a Saturday morning and finishes the following Friday evening during which time we undertake 2500 personnel hours of work on the train.

    “Having started in August 2011 we are now 60 per cent of the way through the programme. At the start of the project, we went through a steep learning curve but now we are into full production where we are turning around one unit each week. There are a lot of refreshed trains in service that our passengers are benefitting from as a result of this overhaul work.”

    Perhaps the biggest challenge has been delivering that work safely and efficiently given the number of people working on and around the train. Justin Lanigan, repair shop manager at Selhurst Depot, explains how this was achieved: “We’ve applied continuous improvement techniques to the process. We’ve also worked to remove unnecessary waste with improved tooling, lineside equipment, and designated areas for materials and storage. In order to manage the work effectively, we introduced a number of strict KPIs and targets, and carried out daily quality audits.”

    Aside from the obvious benefits in terms of time management and cost savings, carrying out the work internally has enabled Southern to learn lessons that can be applied to other projects. “A prime example of this is that we have just started the equivalent overhaul on the Class 171 diesel units, so a lot of the principles and processes we applied for the 377s are being implemented from the start,” notes Iain. “We also have another fleet of trains, the Class 455s, that are being refurbished off-site at an external works so the fleet as a whole is going through quite an uplift at the moment.”

    Innovation is also being applied to some of the areas that passengers don’t necessarily see. This includes the Third Rail In-Service Monitoring Equipment (TRIME) project, which was named as joint winner of the Stephenson Award for Engineering Innovation of the Year at the 2012 National Rail Awards.

    “With TRIME, our objective was to try and get a quantifiable measurement of what is going on between the third rail and the shoe-gear on the train, and to try and understand where the interface between them is worse than expected,” describes Simon Green, Southern’s chief engineer. “We believe there is a whole industry benefit here as a bad shoe-gear incident can cause significant disruption and delays. Therefore there was a clear business case to try and understand where these problems were occurring and then having a plan to deal with them.”

    Together with its partners, Network Rail and Birmingham University, and with support from the RSSB, Southern developed TRIME, which was fitted to one of its Class 377 trains in 2012. Since then the company has continued to tweak and refine the equipment in order to gain further benefits: “We’ve focused on improving the reliability of the laser measurement kit with new mountings, and have started to integrate some track monitoring capability which will hopefully give us some data on ride quality as well,” highlights Simon. “So far we have seen the data generated being used to target maintenance interventions, and although we need to do more validation work, we tend not to see so many precursors of broken shoe-gear frangible joints, which would indicate that we are slowly removing the areas of poor interface performance on the network.”

    Remote condition-based monitoring is already being made use of on the Class 455 fleet through its Train Automatic Performance Analysis System (TAPAS). Developed together with HSBC Rail and Tessella, the system records and downloads journey files that can then be analysed and used to optimise fleet reliability and performance.

    “Because the OTDR system on the train is there predominately to understand what happens from an operational perspective, the question arose as to whether we could use that information to look at how the driver drives the train as a support to driver competency assessments,” says Simon. “This led us to develop the CHURROS system, which analyses journeys by driver rather than by unit. By looking at the data we can discern individual driving style and flag up characteristics that may be considered as precursors to a safety of the line incident.”

    He continues: “This is not intended to be a spy in the cab, but a tool that helps manage the safe operation of our railways, by facilitating a meaningful conversation between the driver and line manager. We are now in discussion with the RSSB to explore the possibility of starting a research project into this technology and how it can be used going forward.”