Machine Learning Driven IoT Enabled Real Time Energy Management Strategies for xEV Charging Systems in DC Microgrids

The integration of electric vehicles (EVs) into DC microgrids powered by renewable energy sources (RES) introduces challenges related to stability, power quality, and anomaly detection. Traditional monitoring methods often struggle to efficiently detect real-time anomalies, resulting in power quality degradation and increased energy losses. This paper presents an IoT-enabled real-time monitoring system combined with a CNN-based anomaly detection framework for xEV charging in DC microgrids. The system uses Wireless Sensor Networks (WSN) to gather real-time data on DC voltage, power demand, and battery state of charge (SoC). This data is then processed using Convolutional Neural Networks (CNNs), which convert time-series sensor readings into structured features for fault classification and predictive analytics. The proposed method enhances power stability, improves fault detection accuracy, and reduces harmonic distortion losses, ensuring an efficient and reliable energy supply for xEV charging stations. MATLAB-based simulations validate the system's effectiveness, demonstrating its ability to detect anomalies, optimize energy distribution, and improve overall power quality in DC microgrids.