Train energy consumption

Discover how we measured train energy consumption with virtual sensors based on a GPS signal and altitude maps, in this project with DB Mindbox.

Transport efficiency as a starting point

Although railway is one of the most energy-efficient transport modes alongside with water transport, monitoring and reducing train energy consumption still makes a lot of sense, for operative transport efficiency and for reducing CO2 emissions in case of diesel locomotives.

Since October 2020, all new and renewed locomotives on EU infrastructures must be equipped with energy measuring equipment. Such equipment helps to understand how much fuel or electricity does a train use per km and allows to:

  • correctly allocate energy use by train / by trip
  • inform the train driver timely about his energy consumption
  • and thus to encourage energy savings.

In some cases, energy management systems allow also billing of energy consumption by rolling stock. Nowadays, all new diesel and electric locomotives are equipped with telematics. However, the average service life of a diesel locomotive is typically 40 years and more, and 30 years for an electric locomotive. This means that there are still a certain number of older trains in operation, which were built before the advent of energy measuring equipment and technology to send and receive data from locomotives on rails to a central server (telematics).

Our Proof of Concept (PoC) with DB Mindbox

In January 2021, COMPREDICT was among the winners of DB Mindbox – the Startup Program of Deutsche Bahn (DB), Europe’s largest rail infrastructure and mobility provider. DB Mindbox is the interface between startups and the DB. The DB Mindbox team matches startups directly with business units of Deutsche Bahn, and so they did with COMPREDICT.

The business unit we were working with was seeking to get a better understanding of the energy consumption of their electric trains, in particular such older trains which are not equipped with any telematics solution for data recording and data transmission.

Train energy consumption | COMPREDICT

How much energy does a regional or high speed electric train use?

Modern high-speed trains are equipped with data recording and telematics software to measure energy consumption.

The question of energy consumption becomes tricky, if a train does not have data recording and telematics yet. Retrofit of modern data loggers or telematics hardware can be extremely costly, these locomotives have not been designed to exchange, store and transmit any data. As the pantograph supplies the locomotive with electric energy from the overhead train line, the energy input can be measured – but not continuously, just over one trip as a whole.

Our customer’s target was to get a much deeper understanding of how trains consume energy, and in a second step to encourage the train drivers to energy-efficient driving.

A virtual sensor for energy consumption, based on GPS

Input data we have used

In our Proof of Concept (PoC), the DB has equipped some of their older electric locomotives with GPS modules. These online navigation data were combined with altimetric data of the examined railroad sections, as inputs for our virtual sensor for energy consumption. Vehicle data of the different locomotives then allowed to adjust the virtual sensor’s parameters to each locomotive model.

Outcome of the PoC

Like in any AI-based project, the quality of the input data revealed to be a key factor for the quality of the results. For some locomotive data, the vehicle data were of very good quality, so that we could get excellent correlation between the software-based instantaneous energy consumption and the hardware-based measurement of the overall energy consumption per trip.

Outlook to the future – energy consumption dashboards

COMPREDICT’s virtual sensor for train energy consumption can help railway operators to optimize all aspects of train energy consumption:

  • acceleration phases
  • up-hill driving
  • brake energy recuperation

Our virtual sensors run on AI-CORE, our machine learning hub. This hub can be interfaced via API/SDK into an existing railway operator’s dashboard. Such dashboards, for example with a color code from green to yellow, orange and red, can help to educate train drivers about potential energy savings. Timely feedback also values their contribution to transport efficiency.

If required, we can also give access to AI-VIEW, our web platform for data upload, results download and graphical results visualization. We are open to provide customized dashboards for the railway sector.

Connected use case: freight train energy consumption

In our PoC we only examined energy consumption of passenger trains. However, the methodology is very easy to adapt to measure the energy consumption of rail freight transport.

Get in touch with us to discuss about your use case!


+49 (0)6151 3844614


Rheinstraße 40-42

64283 Darmstadt, Germany