Francesco Popolizio, Systems and Control

Electric vehicles (EVs) are rapidly becoming a mainstream transport technology, and their growing diffusion is influencing mobility and electricity systems. Since most cars are parked for most of the day, their idle time can offer an opportunity to provide flexibility. When equipped with bidirectional charging capability, an EV can support vehicle-to-home (V2H) operation by supplying household demand, vehicle-to-grid (V2G) operation by exchanging energy with the grid, and increased photovoltaic (PV) self-consumption by storing surplus solar generation. Unlocking this potential, however, requires control strategies that manage uncertainty in household demand and solar generation, satisfy user-driven mobility requirements, and avoid excessive battery wear.
This thesis develops an online framework for residential vehicle–home–grid energy management with rooftop PV integration. During each home-parking interval, energy flows among the EV battery, household load, grid, and PV system are scheduled using a shrinking-horizon model predictive control (SH-MPC) approach that updates decisions as new information becomes available. Battery lifetime effects are modeled by combining calendar and cycle aging mechanisms. To handle imperfect foresight, the controller is coupled with a neural-network-based forecaster to estimate future household load and PV generation. The control objective is to minimize total operating cost, jointly accounting for electricity purchase/sale and battery degradation expenses.
Simulation studies under Swedish residential conditions demonstrate that degradation-aware bidirectional charging can provide tangible economic value while maintaining acceptable battery aging. Sensitivity analyses confirm that these benefits persist across a wide range of operating scenarios and uncertainty levels.