GreenFlux performs smart charging with machine learning

A road sign showing charging sign for electric cars

World’s first implementation of smart charging based on machine learning: GreenFlux, Eneco eMobility and Royal HaskoningDHV achieve a world’s first in EV charging.

Royal HaskoningDHV's office in the city of Amersfoort, the Netherlands, is the first location in the world where electric vehicles are smart charged using machine learning. The charging stations are managed by the charging point operator Eneco eMobility, with smart charging technology provided by the GreenFlux platform.

With the number of electric vehicles ever increasing, so is the pressure to increase the number of charging stations on office premises. This comes at a cost though: electric vehicles require a significant amount of power, which can lead to high investments in electrical installation. With smart charging these costs can be significantly reduced, by ensuring that not all vehicles charge at the same time.

With the innovation developed by GreenFlux, deployed by Eneco eMobility, and applied at Royal HaskoningDHV’s Living Lab Charging Plaza, smart charging is now taken to the next level, allowing up to three times more charging stations on a site than with ‘ordinary’ smart charging.

Machine learning

The novelty in this solution is that machine learning is used to determine or estimate how charge station sites are wired physically, data that commonly is incomplete and unreliable. At Royal HaskoningDHV, the algorithm determines over time the topology of how all the three- phase electricity cables are connected to each individual charge station. Using this topology, the algorithm can optimize between single and three phase charging electric vehicles. Though this may seem like a technicality, it allows up to three times as many charging stations to be installed on the same electrical infrastructure.

Download the whitepaper: EV Smart Charging: Optmising with Renewable Energy

Now that this part has been tested and proven, there is so much more we can add. We can use the same technology to, for instance, predict a driver’s departure time or how much energy they will need. With these kinds of inputs, we can optimize the charging experience even further. — Lennart Verheijen, head of Product Management at GreenFlux:

Living Lab

Mike van Gemund, Project manager ICT and Fleet manager at Royal HaskoningDHV:

'We started with our 100% EV company car policy in 2017. Optimizing charging solutions is key to facilitating our large EV fleet. Given this, we were excited to participate and make our head office’s charging plaza available to Eneco eMobility and GreenFlux as a living lab, providing customer and EV-driver feedback in order to get more out of our current charging infrastructure. Sustainability and innovation are part of our core as an engineering consultancy; we practice what we preach!'


Bart Fick, Lead Technology at Eneco eMobility:

'Currently, the roll-out of smart charging solutions is project-based, however, we aim to make this more universal, and scalable and at the same time extend the effectiveness and flexibility of smart charging. This is required since I expect that in the next years, every new charging station will be equipped with Smart Charging functionality. Since every local installation is different, phase-specific steering is one of the largest and most challenging hurdles for smart charging.

With the GreenFlux/Eneco solution, especially since it is cloud-based and therefore charge-station independent, we now have an easy and reliable solution for this issue, for both new and existing installations.'