Analysis of Constraints for Optimal Electric Vehicle Charging

The increasing uptake of electric vehicles is likely to put a significant demand on the electricity grid. However, expensive infrastructure upgrades can be avoided if some of the vehicle charging can be shifted to off-peak times. We express electric vehicle charging as a receding horizon optimisation problem in which inherent network limitations are taken into account as linear constraints. By tuning these constraints, it is possible to push total system performance in one direction or another as required, and to adapt the solution to a variety of networks and to different sets of network limitations. We explore in detail this tuning process and discuss the trade-offs involved. There are several emergent benefits from examining the constraints in this way: a capacity constraint can be used to prevent peak load increases; a phase unbalance constraint can be used to enforce system rebalancing; and a voltage limit constraint can ensure that vehicle charging does not push voltages outside allowed limits. Our conclusions are demonstrated in simulation studies that use a model of a real network, as well as real demand and vehicle travel data.