All aboard the intelligent train: data-driven optimisation at Dutch Railways

Wim de Liet explains how information from trains is being combined with other data to create a multifaceted real-time picture of the railway

Holland is one of the most densely populated countries in the world. To add more railway tracks to increase capacity is therefore an expensive last resort. Instead, state-owed Dutch Railways (Nederlandse Spoorwegen, NS) is using technology to optimise the process of getting passengers on and off trains, and also to plan for the capacity of new rolling stock, and provide reliable real-time travel information.

Wim de Liet, head of the IT service organisation for travel information for trains, explained that his role combines two areas of focus: being in charge of the IT on NS's own trains such as Wi-Fi and sensors, and also providing travel information for the rail network as a whole, including for private companies' trains.

NS's own rolling stock is being transformed into "intelligent trains, measuring, speed, heat, information from the train control management system, GPS, and other things," he said.

The information from all of the firm's intelligent trains is combined into a "virtual train" in the back office where it can be melded with other information including that coming from other firms' rolling stock and information systems.

One of the main uses for the intelligent train data is improving passenger throughput.

"We can tell how many people are on a train from Wi-Fi and how full each carriage is," de Liet explained. "We are experimenting with intelligent platforms that read information from the train so we can tell people where to stand on the platform so they can get a seat."

He continued: "By getting passengers in and out faster you can put more trains on the track. Instead of four an hour we can go to five or six an hour."

The information from the virtual train is combined with travel information to feed an app, NS Reisplanner, which passengers can use to see trains moving about on a map of Holland (see picture) and to drill down into information about each train, such as arrival time, number of carriages, or the reason for any delay."

Getting the dots to move reliably in real time on the map was surprisingly difficult, de Liet said.

"You have to combine all the travel information and all the logistics information and turn it into little dots on the map. When the train stops the dots have to go, but if there's a delay the dot stays. It's a huge thing, it took us a year to get it right."

NS uses Tibco's BusinessEvents solution to automate the process of collecting the data, analysing it in context and acting on it in real time. A joint tender by Tibco and CGI was accepted over a rival bid by Software AG, even though the latter had run the original proof of concept.

The project has involved standardising all the hardware and software on NS's trains as well as creating the data driven services for operational efficiency and customer experience.

It's still early days. The company has plans to introduce predictive maintenance and to pull in data from other sources, such as weather, to create a more detailed picture. The company is also buying new trains and this will be informed by the data. For de Liet, though, one of the most valuable benefits is that it has proven the value of advanced technology and analytics to the business.

"Before it was very hard to get the business to start this, because every department was involved with the logistics, the finance, the travel information, what the government said we have to deliver," he said.

"But now a lot of people are coming to us to tell them how to get more performance out of their business systems."