Stochastic modeling for scheduling the charging demand of EV in distribution systems using copulas

• We proposed two strategies to pre-schedule the charging of EVs in the next day.
• One of the proposed strategies needs unidirectional communication infrastructure.
• Demand in the next day was appropriately estimated using copula through less data.

Charging demand of electric vehicles (EV) has potentially a significant influence on the power grid. If this charging demand coincides mainly with the peak demand of the power grid, then additional active power has to be supplied to fulfill load management purposes. Thus, it is necessary to estimate and schedule the charging demand of the EV in order to lower the peak demand. Various estimation techniques are available such as Gaussian mixture model and copula. This paper uses copula for data estimation because copula imposes no restriction on the marginal distributions of the available data. Meanwhile, uncertain estimated data requires error elimination. Clayton copula is selected for flexile part of load profile, and Gaussian copula for non-controllable part of the load profile based on the two consecutive days (TCD) classification. Hence, the created scenarios were applied to an optimization problem that flattens the load profile as much as possible using general algebraic modeling system (GAMS). Then, this research concentrates on two new semi-automatically proposals concerned with the day-ahead charging demand response (DR) strategies. These strategies contribute to removing the estimation errors created due to the uncertainties. In order to examine the efficiency of the proposed strategies, the day-ahead pricing (DAP) with inclining block rates (IBR) model is assumed for retail electricity markets.