What to Read Next
As director of Mobility Data Services for Siemens, Gerhard Kress literally helps the trains run on time. Specifically, he and his team of data scientists work to keep European trains and rail schedules operating at the most optimum levels.
As an example, Kress’ team has worked with the Spanish national railway, Renfe, with a goal of ensuring on-time operations on the high-speed rail line between Madrid and Barcelona. The rail trip takes two and a half hours, and it competes with air flights of about an hour and twenty minutes. Renfe upped its ante by promising train passengers a complete refund if the train is delayed by 15 minutes or more. Kress’ team helped Renfe achieve nearly flawless service, with only one major delay in 2,300 trips. Other rail clients for Siemens’ data services include the Bangkok Metro, the Russian rail service between Moscow and St. Petersburg, and trans-European Eurostar service.
Kress’ task in the digital transformation of rail service is to identify the right kinds of data and then use it to manage better. Kress’ team keeps track of trains in real time (including details on mileage, how elements such as brakes are working, the behavior of compressors, the weight of connected rail cars, and the status of automatic control processes), forecasts wear and tear on components (based in part on the quality of the rails, the terrain trains are traveling on, and weather conditions during operation), and — crucially — predicts when breakdowns might occur. The data team builds maintenance scheduling around its data, helping to ensure that service teams are ready with a full array of information and a prescriptive plan for what maintenance to perform when trains arrive at service centers.
Siemens is one of the world’s largest producers of energy-efficient technologies, with customers in power and gas, healthcare, and financial services. Based in Germany, the company focuses on automation and digitalization. As of September 30, 2015, Siemens had around 348,000 employees in more than 200 countries, with revenues in fiscal 2015 of €75.6 billion.
On April 26, 2016, MIT Sloan Management Review hosted a webinar, made possible with sponsorship support from Teradata, with Kress. He talked how his team wrangles so much information — a fleet of 100 modern rail cars produces between 100 to 200 billion data points annually — and what lessons his experience provides for others. The webinar was moderated by Bruce Posner, senior editor of MIT SMR, and highlighted on Twitter at the hashtag #MITSMRevent. Among Kress’ key points:
Transportation mobility is at the center of our society.
Kress said that when you look at, for example, big cities, transportation brings people to and from work, moves large amounts of goods and products. Making transportation work means not just focusing on cars and trams and trains themselves, but on the systems that makes sure that there’s safety, power, and communication around all the movement of every transportation unit. Especially important, says Kress, is rail: It would be hard to have a big city operating efficiently without a functional rail system.
Email Updates on AI, Data & Machine Learning
Get monthly email updates on how artificial intelligence and big data are affecting the development and execution of strategy in organizations.
Please enter a valid email address
Thank you for signing up
Digitization is now a central strategy for supporting rail companies.
Kress said that Siemens’ strategic focus for helping its rail customers with their operations is three fold: Electrification (making sure the train rails have electricity), automation (automating the movement of rails), and digitization (connecting real-world assets to a digital network, creating a digital representation of real-world objects, and supporting real-world processes through digital workflows, providing appropriate feedback).
“One of the promises that we try to achieve is making sure we get 100% operational availability.”
Simply put, that means that a rail vehicle can always be used when it’s scheduled to be used. It’s a big promise with a tangible, clear payoff, Kress said: “This means that passengers that are waiting for a train are getting a train.” How do you make sure this happens? A data team monitors the more than 1 billion data points that a train sends out every year (“modern rail vehicles are very talkative,” Kress said), including information about what parts of the train have broken, what spare parts have been used, what spare parts are still available, what geographic regions it has traveled through (is there a hill that is notorious for causing problems?), and whether it’s near a service depot that has capacity to provide maintenance.
“Data itself is useless unless we have knowledge to turn it into information.”
Even more important, Kress said, is to then take that information and turn it into an action. For the rail industry, that means making decisions about whether to take a train out of service to be repaired or keep it running for a few more days. “Again, you need a lot of knowledge to turn it into information about the most appropriate action that does not disturb operations.”
Siemens had to invest in its own data team before it could help customers establish deeper IT capability within their operations.
Its first experiments, Kress said, were simply based on engineering know-how and didn’t involve special data toolsets. As Siemens began hiring data talent, it established a data lab, using a Teradata UDA (Unified Data Architecture) system that could handle a mass volume of information. At that point, Kress said, “we spent a little bit of time trying to figure out what kinds of insights we could get from the data.” This was important, Kress said: “One of the issues I have discovered in organizations in the past is that there’s a great idea about data and its value, but you need to prove it.” Initial investment can be high, which means there’s pressure to show a business case for the expenditure. With a few good cases under its collective belt, Kress was able to ramp up the data services team to 30 people, solely focused on data analytics and services.
Data services in the rail market currently focus on four key value levers.
Those levers, Kress said, are improved maintenance, root-cause analysis for failures, reduction of preventative maintenance cost (by not replacing things that don’t actually need to be replaced yet), and predictive maintenance. “The core element of predictive maintenance really is to make sure you don’t have a failure of a vehicle somewhere on a journey, but you understand the upcoming problem before, and you can replace the component at the time when it suits you best.” For example, that means that for high-speed trains running in Europe, Kress’ team can predict the failure of a gear box 7-10 days in advance. That gear box can then be replaced at a service depot “somewhere at night, when there are no passengers on the train.” The goals, he said, are improving availability of trains, and “making sure that there are no surprises in the operations.”
Not only do data-driven operations provide more dependable service, but they have a hidden savings: companies can operate with less spare capacity.
Having maintenance managed by sophisticated analytics means that companies don’t have to buy and maintain excess rail inventory to ensure that there are always trains ready to go. “What we see a lot across the world is that a lot of operators have up to 10-15% spare capacity in their vehicle fleets,” Kress noted. “If you use data in this way, and you really utilize all the predictions, then it’s possible to operate the same network with much, much less spare capacity.” Kress said that one company that Siemens works with, in Bangkok, actually runs its system with no spare capacity.
The more reliable rail services are, the more competitive they are with non-rail options.
In Spain, Kress’ team worked with Renfe, the national railway, to boost on-time operations on the high-speed rail line between Madrid and Barcelona. The rail trip takes two and a half hours, and it competes with air flights of an hour and twenty minutes. When the trains started, 80% of travelers took the plane, and 20% took the train. As of early 2016, those number, Kress said, have reversed: 80% are taking the train, and just 20% are flying. “That is because those trains are punctual,” said Kress. “That is great. And this really has helped Renfe operate a very good, very profitable service.”