Partial privatisation studies included in DB restructuring
The Next Federal Election and Passenger and Freight Rail
Transport and Logistics symposium to gauge railway link
Rail gets another CRC. Third time lucky?
Chinese high speed rail should confine the XPT to history
Hendy heads to NR
Urban rail news in brief - July 2015
Inland rail a trifecta for Toowoomba region: mayor
The answer to the Great Question of Life, the Universe and Everything is 42. After seven-and-a-half million years of calculations, this was the conclusion delivered by Deep Thought, a supercomputer in Douglas Adams’ novel The Hitchhiker’s Guide to the Galaxy. When Deep Thought’s hyper-intelligent pan-dimensional creators contest her findings, she replies, “I think the problem, to be quite honest with you, is that you’ve never actually known what the question was.”
This visionary tale was written long before the advent of Big Data, but it carries a message pertinent to the modern railway: Data in isolation has no meaning, and the real value lies in how it is used. Infrastructure managers and train operators harvest increasing volumes of data about their assets, and this is the starting point for potentially transformative insights that could have profound insights for the way railways operate. But in a world where assets are becoming more and more digitally interconnected, extracting meaningful insight from the terabytes of data streaming in from billions of data points on trains, track and signalling systems is a hugely challenging task. Digitalizing asset management and digitalizing assets must therefore move forward in unison if the potential of data is to be unlocked.
“If rail wants to survive competition with driverless cars, we need to fundamentally change the way we operate, and this includes moving to CBM,” Gerhard Kress, Director of Mobility Data Services for Siemens, told conference delegates. “It’s essential to focus on creating value and not just harvesting huge amounts of data, because data alone is only cost. The real benefit comes when you can turn it into insights. Everything we do around data has no value unless it results in tangible benefits for the customer.”
According to Matthew Miller, Global Transportation Industry Principal for OSIsoft, market hype around the potential for Big Data is being driven by four megatrends:
Miller says convergence of these key enablers is driving the adoption of IoT (Internet of Things) in rail, and with it demand for analytic applications to turn data into operational intelligence for better decision-making. However, he warns that organizations need to be able to cope with variation in data quality. “Real data isn’t perfect, and data quality is a killer issue for every project,” he says. “Communications failures, spiking sensors, calibration issues and out-of-sequence data can all cause massive problems, and you need to consider how you will deal with these issues.”
Another challenge is the increasing volume of data being generated. As IoT expands, the quantity of data from connected devices looks set to rise dramatically. Even with the advent of 5G, centralized processing of such massive volumes of data will place a huge strain on telecommunications networks. Edge computing—moving the processing of data closer to the IoT devices that created it—is currently a focus for IoT firms around the globe.
“Analytics is a challenge,” says Michael Thiel, CEO of Frauscher Sensor Technology. “With artificial intelligence (AI), a lot will be possible, but where should we place the analytics for that? A sensor network monitoring 40km of track can generate a terabyte of data in an hour, so we need to reduce it to the point where it can be handled easily, and that’s what we’re working on now, shifting intelligence closer to the track.”
Condition-based maintenance (CBM) is a key business driver for digitalization in the rail sector. According to The Rail Sector’s Changing Maintenance Game, a report published by McKinsey & Co. in December 2017, CBM can reduce rolling stock manual diagnostics by at least 60%, and could lead to an overall reduction of at least 10-15% in maintenance costs—equivalent to an annual saving of up to $4.7 billion for train operators, $2.35 billion for rolling stock OEMs, and $4.7 billion for third parties.
According to Perpetuum, which develops asset management solutions for rolling stock, 9-12% of total vehicle operating costs are spent on truck maintenance, and lifecycles can be extended significantly with the aid of remote condition monitoring (RCM). “We are burning millions of [dollars] replacing assets that are not life-expired,” explains Perpetuum Global Sales Director Robert Mulder. “If the condition of every single bogie was known, we would be able to extend overhaul intervals by 25-75%. With vibration, the start of degradation is already picked up 6-7 months before the asset is replaced.”
Perpetuum has developed a “fit-and-forget solution,” which is self-powered by vibration and has a “maintenance-free” design life of 20 years. The technology enables remote monitoring of wheels, bearings, brakes, axle boxes, traction motors, and track quality in real time. Bespoke algorithms in the cloud provide the operator with real-time status updates.
“You have to do more than simply implement RCM,” says Mulder. “To gain the potential benefits, processes need to be in place. No technology project is going to generate a return on its own.”
For OEMs looking to build their aftermarket business, asset management has become a key focus in recent years, and suppliers are helping their customers to bridge the gap between raw data and meaningful insights into asset status and performance. Siemens uses its Railigent platform to remotely monitor gearboxes, bearings and traction motors on Velaro high-speed trains, as well 5,000 doors on the fleet of Desiro City EMUs used on London’s Thameslink network. Siemens has also used machine learning to predict bearing failure on high-speed trains and high-end data analytics to predict point machine failure without the need for additional sensors.
Siemens is working with third-party suppliers to integrate their applications into the Railigent platform. Voith recently signed an agreement with Siemens to develop a monitoring solution for its Scharfenberg coupling, and Siemens is also working with SKF to integrate its Insight Rail CBM solution with Railigent to optimize bearing maintenance.
According to McKinsey, railway companies should be looking to initiate three short-term steps to begin preparing for CBM and predictive maintenance:
McKinsey also warns that companies should not underestimate the organizational challenges of cross-department or cross-company collaboration, culture clashes between data analysts and railway engineers, and the transformation of maintenance processes.
Dr. Burkhard Schulte-Werning, head of the maintenance business line for German Rail (DB), told delegates that optimization of conventional rolling stock maintenance practices is reaching its limits, and DB’s vision is to use continuous data transmission to link its assets to its InfraView open data platform as basis for CBM.
In a step toward this goal, DB Cargo will equip 2,000 locomotives of 16 different types with telematics systems by 2020. DB Cargo’s TechLok system collects data from the locomotive, with up to 7,000 different diagnosis codes or status messages for each type. Data is transmitted in near real-time via a GSM connection for visualization and analytics, which is used to develop use cases and transform data into usable information and results. DB Cargo is using GE’s RailConnect 360 asset performance software as well as Siemens’ Railigent platform and MindSphere IoT system.
DB is also using scheduled trains for continuous track monitoring (CTM), with around 2,500 km of track now under CTM supervision using an ICE 2 high-speed train, a class 189 electric locomotive, and a class 424 EMU on the Hanover S-Bahn network.
A $1.8 million research project has been launched as part of the Shift2Rail Joint Technology Initiative, which aims to optimize rail vehicle maintenance processes through the integration of predictive data analysis algorithms and online optimization tools within a CBM strategy.
The maintenance element of the Smart Maintenance and the Rail Traveler Experience (SMaRTE) project will focus on the use of information and modelling to reduce lifecycle costs and improve vehicle availability and performance through CBM.
The 28-month project is due to conclude in December 2019.
In 2014, Belgian rail freight operator Lineas began experimenting with asset data with the aim of generating a competitive advantage. One of its most successful initiatives was optimizing maintenance planning for its fleet of 110 class 77 diesel locomotives, which are assigned to a variety of duties from shunting to longer-distance main line operations. The IoT solution provides data every 10 seconds, and Lineas developed a dashboard to provide a complete overview of the fleet’s status. “Small and medium-sized companies were brought in to carry out analytics on the fleet and explain the problems to our bosses,” says Director of Assets and Network operations Jeroen Spruyt. “We constructed a Know Your Locomotive view, which was used to build a predictive model of failures. This will lead to a reduction in capex of $1.2 million on a total annual fleet running cost of $29.4 million and will also stretch the life of the locomotives.
Lineas has used Big Data to optimize maintenance planning on its fleet of class 77 diesel locomotives. Quintus Vosman photo.
“We challenge people in our organization to come up with a business case for small projects. For Lineas, this is no longer a hobby. In 2014, we were playing around; today we’re building an IoT backbone and equipping our fleet with IoT solutions. To solve the questions of the future we need to start building models today.”
Netherlands Railway (NS) has established a competence center for advanced analytics that works closely with analytics teams in all other departments of the company to develop dashboards, data flows for use in business processes, and products that will benefit end users. Through this cross-department collaboration, NS has developed a model for monitoring air leakage from train brake and door systems, which reduces the need for an operator to walk alongside the train.
Another innovation is the Zitplaatszoeker seat finder app currently being tested on the Arnhem – Nijmegen – Den Bosch line. The app uses a color-coding system to display levels of seat occupancy throughout the train, helping customers to find a seat and ensuring a more even distribution of passengers. The seat searcher app uses data generated by weight sensors on the track that were originally installed to weigh freight trains as part of a plan for weight-based track access charges.
Netherlands Railways Zitplaatszoeker seat finder app.
Austrian Federal Railways (ÖBB) has harnessed Big Data and the company’s asset management system to support the development of its Target 2025+ long-term infrastructure strategy. ÖBB Infrastructure Senior Asset Manager Richard Mair told delegates that clearly defining desired outcomes at the beginning of the process was critical to successful interrogation of the data.
“The objectives [of the strategy] were so big that they needed to be broken down,” he explained. “We had a lot of ambiguous data, so we needed an Asset Management System (AMS) as a single point of truth providing us with a good description of the base and supporting a vision for the future. The AMS allows us to answer the questions of asset owner, and there must be a question, otherwise you cannot provide an answer. Big Data is not a magic wand: Before you start to search for a solution, you need to know what outcomes you want.”
For ÖBB, the AMS was the enabler between raw data and business insight. “Information is one of our most important assets today and it will be even more important in the future,” Mair said. “However, data doesn’t help you at all with information. You need IT systems to aggregate and present data in a meaningful way.”
In February 2017, Dutch infrastructure manager Prorail established DataLab, which harnesses Big Data to develop solutions to issues affecting the performance of the network, including switch and crossing failures, track defects, signalling and train detection faults, trespassers, and stray animals.
Predictive models are developed in four-week scrum cycles by a team that includes experts from relevant areas of the business with input from academic institutions, contractors, engineering firms and tech startups.
One of the first DataLab projects involved developing a predictive model for trespassing, a common cause of operational disruption on the Dutch network. This looked at key influencers—environment, hotspots, weather, school holidays, and ease of access to railway property—and used machine learning to develop a predictive dashboard that could predict the risk of trespass at key locations. Police have been using the dashboard since August 2017, and trespassing incidents have fallen by 50-100% at the locations covered by the system, with a doubling of the arrest rate.
The DataLab has also developed a predictive model for switch failures that is capable of detecting up to 20% of faults before they occur. Last year, Prorail awarded machine-to-machine technology company Dual Inventive a contract to supply 1,500 wireless sensors to remotely monitor the health of points and crossings, and a further 500 sensors to monitor other infrastructure systems such as level crossings.
Belgian infrastructure manager Infrabel has also embraced big data and built an analytics platform to find answers to key questions about its assets. Infrabel has developed a model to evaluate how frequently trains approach red signals in a bid to reduce Signal Passed at Danger (SPAD) incidents. The resulting model integrates data on signal aspects, train movement, infrastructure and track occupancy to identify red signal hotspots and build a SPAD risk index model. Big data has also been used to support a program to reduce the number of switches and crossings on the network, which is intended to improve reliability and cut maintenance costs.
French National Railways (SNCF) has begun fleet rollout of an IoT-based solution for monitoring the status of passenger train doors in France. The Avisé system has been developed as part of the Digital SNCF program to automatically notify traincrew if the doors of Corail coaches are not properly secured when the train is in motion.
Two sensors remotely monitor the status of each door, illuminating a lamp in a vestibule cabinet if the door is properly secured. The status of the lamp is transmitted via a communications module to SNCF’s IoT platform, which compares the position of all doors of the train and issues a smartphone alert to the train crew if it detects an anomaly. This enables unsecured doors to be quickly identified.
SNCF plans to install Avisé on 350 Corail vehicles, which are used on Intercités services across the country.
Swiss Federal Railways (SBB) has developed the Swiss Track Analysis & Maintenance Planning (swissTAMP) tool, which integrates data from various sources to provide information on asset condition, which in turn supports decision-making on maintenance measures. This means planning for infrastructure maintenance and renewals can be carried out in a traceable and needs-driven manner. The tool enables visualiZation and analysis of component and system data and provides site-specific prognoses of future maintenance requirements. SBB says swissTAMP is playing a key role in the transition toward preventative maintenance on the Swiss rail network.
All these initiatives demonstrate the central role of asset management systems in turning data into insight. Organizational factors are also key: Success depends on a company’s ability to fuse traditional railway disciplines with data science and adapt to new ways of working. Most important, Big Data can only provide an answer if there is a question, and a clear focus on the desired outcome is essential if the transformational potential of Big Data is to be realized.
The post Bridging the Big Data gap appeared first on Railway Age.
This article first appeared on www.railwayage.com
About this website
Railpage version 3.10.0.0037
All logos and trademarks in this site are property of their respective owner. The comments are property of their posters, all the rest is © 2003-2019 Interactive Omnimedia Pty Ltd.
You can syndicate our news using one of the RSS feeds.