FKFS Veranstaltungen

2024 Stuttgart International Symposium
on Automotive and Engine Technology

2. - 3. Juli 2024

Session: CHARGING | | 08:30 - 09:00

Improving Battery Lifespan and Service Trip Reliability of EVs in Public Transport by Learning Energy Consumption

Hermann von Kleist, Fraunhofer IVI

Electrification of public transport in cities puts lots of stress onto the vehicle's traction batteries and the power grid during charging. A self-learning operating strategy (SELEOS) improves the battery life and reduces stress on the power grid by limiting the charging power as much as feasible and avoiding extreme SoC. During regular service operation, SELEOS observes the vehicle state and energy flows inside of the vehicle and between vehicle and charging infrastructure. Based on these observations, SELEOS plans an optimal SoC trajectory for the trip and dispatches recommendations for charging and discharging the traction battery to the vehicle's ECU. Originating from hybrid buses, SELEOS is designed to work on battery-electric buses (BEB) and battery trolley buses (BTB) as well. In addition to reducing stress to battery and power grid, SELEOS ensures reliable service trips. It does so by checking if the current SoC matches the estimated energy consumption in the near future while considering planned charging operations. A situation where the current service trip is endangered can be recognized and mitigated early. In such a case, SELEOS recommends measures which reduce the energy consumption while compromising passenger comfort as little as possible This software can operate offline on vehicles without an active internet connection. However, the learning procedure of the individual vehicles can be accelerated by running the software on a cloud server with a live connection to each vehicle of the same fleet. This way, the online learning software enables similar vehicles of the fleet to 'learn from each other'. Having the SELEOS software on the internet will allow to integrate more information to the planning algorithms in the future. For example, constraints or recommendations from the local power grid provider as well as current energy prices may be incorporated into a charging schedule to reduce further stress on components and operational costs.