Optimal placement of parking of electric vehicles in smart grids, considering their active capacity

In this paper, two important problems of Electric Vehicles (EVs) are simultaneously and comprehensively examined and solved. In the first step, the effects of various exploitation strategies of EVs on the location of commercial and residential parking lots. In addition, the power losses and voltage profile of the distribution network are examined. The optimal placement is performed by considering the load probabilistic model, the parking lots' generation, and a suitable estimation of the annual-hourly load. This optimization problem is solved based on a new mixed approach, including the Cuckoo Optimization Algorithm (COA) with a sequential Monte Carlo simulation. To show the effectiveness of the proposed models, three parking lots are evaluated. In the second step, the performance of the EVs in smart grids is forecasted, which can be used in estimating the grid load and allocating local reserves. The energy injection/charging into/from smart grids by the EVs can greatly affect the balancing/unbalancing of power systems. Therefore, big data for its performance estimation will be needed. Moreover, due to the high volume of these required data in the input vectors, time series and simple neural networks cannot be used. Thus, this problem needs methods that forecast their injection/charging based on their probabilistic contents. In this paper, a probabilistic neural network called the Bayesian neural network is used. This model could rate the probable data based on input data by providing the normal and abnormal probabilistic functions for each hour. The results show the effectiveness of the proposed method based on ac-curacy and forecasting speed.