Identifying Unusual Charging Patterns of Electric Vehicles using Artificial Intelligence

Some of the most challenging parts of finding new approaches to lowering residential energy use involve studying, detecting, and visualizing households' anomalous power usage patterns. This research presents a novel method for identifying irregularities in electric vehicle energy use by extracting features using a modified long-short-term memory model. The latter is implemented to extract load features by analysing intent-driven user consumption instances occurring throughout the day. In addition to feature extraction, we explore the use of a deep neural network, specifically the LSTM architecture, to efficiently detect and classify anomalies in Electric Vehicles. In the following, we provide a novel anomaly visualisation technique based on a scatter representation of the classes, which gives customers a simple way to comprehend unusual actions. These encouraging findings validate the effectiveness of the suggested deep learning approach for identifying abnormal energy usage, encouraging energy-efficient behaviour, and cutting down on energy waste.