DIVERGENT

Parked electric vehicles (EVs) represent a promising resource for storing renewable energy. Smart Bidirectional Charging (SBC) – the integration of EVs with charging stations and local energy networks through data exchange – enables local control of energy flows. This approach leverages vehicle batteries to reduce power peaks at the consumer level and maximize the utilization of surplus renewable energy.

The DIVERGENT project explores methodologies for decentralized decision-making to optimize bidirectional EV charging systems. Our research integrates both traditional algorithmic approaches and cutting-edge machine learning techniques, with a particular focus on multi-agent reinforcement learning (MARL). To ensure a holistic framework, the project also addresses vehicle usage simulations, EV user behavior modeling, and energy management, effectively incorporating key stakeholder requirements.

The practical outcome of DIVERGENT includes the design and evaluation of an AC (Alternating current)-SBC laboratory demonstrator. This system replicates bidirectional energy and data flows between EV batteries, OnBoard Chargers (OBC), automated AC charging technologies, and local household grids.

Key deliverables of DIVERGENT:

• DM methods that support dynamic control of the EV bidirectional power exchange profile and subsequently:
  • Optimize the energy consumption of households/buildings equipped with bidirectional chargers.
  • Avoid peak loads caused by simultaneous charging and discharging of EVs.
• Methods for estimating battery health status and predicting the remaining battery life, enabling the assessment and prediction of the impact of bidirectional charging processes on the battery.
• A laboratory demonstrator for the integration and evaluation of software prototypes for (a) the Electric Vehicle Communication Controller (EVCC) and (b) the Supply Equipment Communication Controller (SECC).