Charging scheduling of plug - in electric vehicles under driving behavior uncertainty

With the perspective of a future widespread penetration of plug-in electric vehicles (PEVs) imminent challenges arise. Since an uncontrolled way of vehicle charging might impose additional system strain by adding up on existing peak-demands, there exists the necessity to find a charging management procedure, which is beneficial in terms of grid operation and cost. In this thesis a centralized control scheme is applied, which hands the responsibility of charging management to a central entity. This entity pools a fleet of PEVs and plans a cost optimal day-ahead charging schedule subject to the aggregated vehicle constraints. This day-ahead schedule then represents the target schedule for the consequent step, the real-time dispatch, in which the pre-planned power schedule is distributed to the individual vehicles while meeting their respective constraints, i.e. energy requirements for the respective driving patterns. One major challenge in the task of finding the best suitable schedules in day-ahead and real-time management is the uncertainty related to the driving behavior of PEV owners, i.e. differences between predicted and actual driving behavior. Therefore, a model of these uncertainties is incorporated in the investigations and a stochastic method is applied to account for those uncertainties in the day-ahead and real-time dispatch accordingly. However, even with the most sophisticated approach for distributing the pre-scheduled power in real time, the existence of uncertainty leads to the issue that the day-ahead schedule cannot be followed at all times. To face this, a model predictive control approach is applied additionally to the predescribed framework to anticipate violations and adjust the day-ahead schedule accordingly. The results show an improved matching quality between real-time and day-ahead schedule with only marginal increase in total power dispatch compared to the real-time dispatch without day-ahead schedule adjustments.