NS turns to AI to cut train electricity use as Dutch power grid is overloaded
Dutch railway operator NS is developing an AI system to advise train drivers on when to accelerate and coast, reducing electricity use on the overloaded national power grid.
NS consumes 4.6 million gigajoules of electricity per year — 1 percent of the Netherlands’ total electricity use and the same amount as the city of Amsterdam, NOS reported. The company has already cut its electricity consumption by nearly a quarter over the past decade. It now aims for an additional 5 to 6 percent reduction, which would equal the annual electricity use of Leeuwarden.
NS is gathering detailed data on how trains accelerate and brake. The planned AI model, expected by the end of this year, will give ride-specific advice by combining real-time train weight (including the number of passengers), track gradients, curves, friction factors, weather conditions, wind direction, and seasonal air density.
Train driver Martijn de Jong demonstrated the technique on an intercity route from Den Bosch to Utrecht Centraal. Several kilometers before Houten, near the Culemborg bridge—where the grid is so congested that new connections are impossible—the train stopped drawing power and coasted the rest of the way.
“Look,” de Jong told a reporter from NOS, pointing to a screen in the cab. “We are now rolling at 110 kilometers per hour through Houten, but we are using no electricity anymore.”
“All the electricity that we don’t have to use can nicely go to the households,” he added after arriving on platform 7B in Utrecht.
De Jong explained the data behind the system: “We have the weight of the train, also via the number of passengers that are on board. Because more weight pushes the train forward. That we combine with information about the track, whether you go up or down, and which curves there are. Everything that causes friction.”
Drivers already use a tablet in the cab, jokingly called the “TimTim,” that shows real-time electricity consumption. The AI will add specific advice to the device.
“The strategy is to accelerate as quickly as possible at departure, to then stay at as high a speed as possible for as long as possible,” de Jong told NOS. “That way, you can coast with the train earlier at the end of the ride. And use as little energy as possible.”
NS says expanding the power grid takes longer than achieving savings through more efficient operations, especially with high energy costs and grid constraints.
