Electric vehicle batteries are becoming more powerful and have more autonomy but also need proper thermal management. Current strategies only take into account the current temperature of the vehicle components to decide when to cool or heat them. The development of smarter decision support systems, such as machine learning models, will lead to better efficiency and consumption criteria.

The focus of this project is to use information about the vehicle’s journey, such as the slope or the speed of the road, to decide when to activate the battery cooling and thus be able to improve current strategies.

Why should a vehicle invest energy in cooling its battery if there is soon a long slope in which it can cool on its own? If it is known that a significant acceleration is approaching, wouldn't it be better to prepare the vehicle in advance to mitigate the effort that it will have to make? These, among other things, are some of the questions that this project seeks to answer.

In conclusion, if we consider not only the temperature of the components of the vehicle, but also their condition, the environment and the journey, we can optimise the performance and consumption of the battery in order to increase its autonomy and lifespan, and minimise environmental impact. The results of the Predictive eBoost project will thus be an important step forward in the implementation of electric vehicles, as they will allow to extend the distance travelled without having to stop, while providing a more satisfying driving experience.

The project is led by Dr Gerhard Lux, of SEAT, and Marc Duevel, of Volkswagen Group Innovation. At the Universitat Politècnica de Catalunya · BarcelonaTech (UPC) it is led by professor Ernest Teniente, the director of inLab FIB, the innovation and research laboratory of the UPC’s Barcelona School of Informatics (FIB).