Electric vehicle charging sessions generator based on clustered driver behaviors.

The increasing penetration rate of electric vehicles (EV) requires the installation of new charge points, which can induce various problems, such as higher peak powers. To address these problems, a well-known approach is to use simulations to assess the impact of uncoordinated charging on a particular
local energy system. Such simulations require EV charging sessions data which are often not (sufficiently) available. This paper proposes a methodology that generates EV charging sessions based on statistical parameters of different type of EV drivers, which have been extracted from historical data via data mining techniques. The results show the great ability of the methodology to generate representative charging profiles for different types of drivers. Additional scenarios are simulated to show the strong impacts of uncoordinated charging for the use case of a hospital.