Optimization of Probabilistic EV Fleet Integration in Unbalanced Distribution System

Electrification of the transportation sector through electric vehicles (EVs) is promoted by environmentalists and government agencies in order to encourage sustainable growth. EVs as a random load may take a toll on the stability and reliability of the power system. However, they can also help to improve the grid performance if operated in coordination with the load profile. In this paper, the impact of EV fleet integration in the unbalanced distribution network has been observed. A probabilistic EV model for charging and discharging of EVs is proposed, in which both the arrival and departure time of EVs is modelled as a normal distribution and the distance travelled is modelled as a lognormal distribution taking into account the spatial temporal features of EV charging. The charging-discharging schedule of EVs is optimized using Genetic Algorithm (GA) with the aim of obtaining a flattened load profile. The developed algorithm was tested on IEEE 13-bus unbalanced test distribution network, and the results show that the optimized EV integration has resulted in flattening the load profile (gap between maximum and minimum demand reduced to ~ 18 kW), improvement in the voltage profile and reduction in the network unbalance as depicted by the decrease in the maximum neutral current drawn (~ 28%). The proposed model can be implemented for practical distribution system planning and can be an effective tool in balancing the unbalanced network in the era of rapidly increasing EVs in the real-time distribution system.