05.02.19
TrackWater: Embracing data in flood prevention
Source: RTM Dec/Jan 19
Mike Harding, from Lancaster University’s Data Science Institute, introduces us to TrackWater: an innovative project between Lancaster University and Transport Systems Catapult.
Network Rail estimates that, between 2008 and 2018, more than three million ‘delay minutes’ were caused as a direct result of track flooding. This has cost Network Rail £119m in service contract fines paid out to private TOCs.
These delays, and other wider impacts of flooding felt by the UK’s growing rail passenger population (1.2 million journeys in 2008 to 1.7 million in 2017), has been caused primarily by external phenomena as the UK experiences more frequent and intense weather events. As well as other infrastructure damage, the summer floods of 2007 and Storm Desmond in 2015 caused widespread disruption to rail services.
This greater unpredictability in our weather, combined with the very nature of rail infrastructure (where a large proportion of lines reside on flat, low-lying areas) makes rail travel particularly vulnerable to even moderate rainfall.
Traditionally, Network Rail has relied on a complex labyrinth of Victorian-built track-side drainage, including catch-pits and culverts, to remove water from the network.
Upkeep of these drainage assets is mainly supported through corrective maintenance and cleansing blockages after flood incidents have occurred. This means cleansing is expensive and reactive at best.
Maintainers have, in part, been constrained to performing corrective maintenance due to limitations in budget. These constraints are also compounded by a limited understanding of where drainage assets are on the network, and the way in which they perform and behave over time.
More recently, Network Rail has taken action towards addressing the growing challenge of flooding on the network. In 2015, a centralised drainage division was introduced alongside more recent work to explore new preventative maintenance approaches to address the problem of surface-water management across the network.
Network Rail has, through the Innovate UK ‘Innovation in Rail’ funding programme, begun to collaborate with an interdisciplinary team of academic researchers, transport practitioners, and technology providers to explore the novel application of emerging internet-connected sensor technology known as the Internet of Things (IoT) and the use of real-time predictive analytics to better inform drainage maintenance decision-making.
A partnership comprising the Data Science Institute at Lancaster University, Transport Systems Catapult, and technology innovator InTouch Ltd has been working together as part of the £1m-funded TrackWater project. The project aims to investigate the feasibility of a new, data-driven drainage and flood management solution, originally developed within the highways sector.
For highways maintainers, the SmartWater system, developed by the same partnership and now managed and delivered by InTouch, created significant cost savings by facilitating a shift to a data-driven approach, using predictive analytics and IoT sensors to provide detailed information so that authorities can tackle on-road flooding proactively.
With an initial focus on the management of trackside catch-pits that often require the removal of silt to ensure flow capacity and mitigate the risk of pluvial flooding, TrackWater’s data-driven approach represents a step-change in how drainage asset service conditions can be inspected, monitored, and maintained.
Through the continuous capture, analysis and reporting of asset conditions in real time, the collection of a catch-pit’s condition is supported through a flexible sensing probe that can be adjusted to the dimensions of an asset to sample relative silt, water, and light levels.
These probes can periodically communicate through available wireless communication frequencies including UHF and an emerging low-power, long-range protocol designed specifically for supporting IoT sensors with a trackside wireless receiver, that further relay this data to a cloud processing environment over the 4G network.
Exploiting these new forms of drainage data, in combination with structural asset information and environmental datasets (such as localised rainfall forecasts), we have developed a series of interpretative and predictive data models. This data provides maintainers with a deeper understanding of drainage behaviour through new descriptive and predictive analytics. These can be applied remotely to inform key business questions such as whether the asset is currently in a state of flood, or understand what the future flood risk of an asset is over the coming week.
Integrating these emerging analytics capabilities within a complex hierarchy of drainage decision-makers, from in-field maintenance teams and route asset engineers to strategic planners at a central level, has required significant engagement.
A diverse range of tools have been developed which integrate the new data feeds and help support more efficient and proactive maintenance decision making.
An end-to-end IoT approach to preventative drainage maintenance has the potential to benefit Network Rail in a number of key areas. Similar to the operational cost savings demonstrated through our previous work within the highways sector, a data-driven approach can help to assist managing authorities in transitioning from predominantly reactive to more proactive drainage maintenance. This frees up limited resource to target problem drainage assets, enabling maintainers to mitigate the risk of flooding ahead of time. In addition, access to historic, real-time and predictive drainage performance information empowers maintainers with a broader awareness of drainage network behaviour, and the way in which conditions change – particularly at known flood hotspots.
Interviews with Network Rail have highlighted potential benefits unique to the rail sector, such as the ability to utilise high-resolution drainage information to guide requests for track possession time and inform cyclical asset inspection intervals.
The InTouch TrackWater prototype system was set up at Network Railʼs test-track in Melton Mowbray in Leicestershire in October. Tests will run until March, when results will be analysed by the team at Lancaster University. Further tests will be run if necessary before the product is rolled out across the UK.
Top image: Andrew Matthews via PA
Enjoying RTM? Subscribe here to receive our weekly news updates or click here to receive a copy of the magazine!