Plug-in electric vehicles demand modeling in smart grids: a deep learning-based approach: wip abstract

In smart grids, Plug-in Electric Vehicles (PEVs) are considered components of the power demand. PEVs have highly stochastic behavior, and to manage this stochastic load efficiently, intermediary bodies, widely known as aggregators, have been developed in the literature. In order to handle the PEVs charging demand from both technical and financial points of view, aggregators include tools based on Internet-of-Things (IoT) technology, which can observe the users' historical behavior and estimate their travel behavior and the requested charging demand. In the near future, the increase in the share of PEV adoption will transform the PEVs demand modeling framework into a "big-data" Cyber-Physical System (CPS). We present a novel artificial intelligence approach based on the deep learning concept to tackle this large dimension problem. To investigate users' different behavior, the PEVs are classified into different groups based on their driving patterns. Then, each class is assigned to its respective deep convolutional neural networks. The proposed method's performance will be investigated in the day-ahead energy market.