The increasing adoption of Electric Vehicles (EVs) has increased the need for an efficient management of charging infrastructure. Load forecasting is one of the most prominent challenges in EV operation and is an important factor for achieving effective scheduling. Various techniques have been deployed for forecasting the load demand in EVs. However, in comparison to traditional loads, the charging and scheduling of EVs is different because of its dynamic fluctuations and periodic variation. These issues affect the performance of conventional load forecasting techniques. The emergence of Deep Learning (DL) models provides a potential solution for addressing the drawback of conventional forecasting methods. Because of its excellent learning ability and capacity to handle large-scale datasets, DL models are extensively used to perform forecasting tasks. This research analyzes the application of different DL models for forecasting the load demand and scheduling in EVs. In this work, various models such as Convolutional Neural Networks (CNN), Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM) and hybrid models are deployed for accurately predicting the load. In addition, a dynamic pricing algorithm is implemented for achieving effective scheduling. A Graphical User Interface (GUI) is designed for verifying charging scheduling and management. Simulation results of different charging stages validate the effectiveness of the proposed framework.
Onteru et al. (Fri,) studied this question.