AI Based EV Charging Station And Scheduling For Smart Energy Systems
Loading...
Files
Date
2026
Journal Title
Journal ISSN
Volume Title
Publisher
NHCE
Abstract
The project presents a refined and fully integrated AI-enabled energy management and EV charging optimization system designed to intelligently coordinate charging operations while ensuring stable and efficient energy distribution within a campus microgrid. The work responds to a critical contemporary challenge: the rapid expansion of electric vehicle usage is placing unprecedented stress on institutional electrical infrastructures, which often lack the flexibility to manage sudden load fluctuations, integrate renewable resources effectively, and maintain operational reliability without costly upgrades [1][2]. To address these needs, the proposed system adopts a Genetic Algorithm-based optimization engine capable of resolving multiple competing objectives simultaneously. These include minimizing electricity cost by incorporating time-of-use pricing signals, lowering peak electrical demand through strategic load shifting, increasing renewable energy penetration by synchronizing charging activities with solar generation forecasts, safeguarding priority loads during constrained supply conditions, and enhancing user satisfaction through transparent and predictable scheduling operations. The optimization module executes at two-hour intervals with a computational time between 45 and 60 seconds, enabling continuous adaptation to evolving load behavior, energy market variations, and environmental factors such as solar irradiance patterns [3][11]. The platform is supported by a wide network of IoT-based sensing and control hardware deployed throughout a 20-hectare academic campus. ESP32-based EV charging units measure voltage and current with ±1.5% accuracy, while Arduino UNO controllers monitor departmental consumption, execute automated load-shedding routines, and analyze three-phase power parameters. Renewable energy nodes incorporating NodeMCU modules facilitate communication with solar and battery storage subsystems, forming a cohesive and distributed sensing architecture. All hardware communicates through an MQTT-based WiFi backbone delivering an average latency of 254.6 milliseconds—substantially lower than the 2-second responsiveness requirement for real-time microgrid management [15]. System-level processing and optimization occur on a centralized MATLAB workstation equipped with an Intel Core i7-12700 processor and 32 GB RAM, maintaining more than 99% availability under continuous operation. Extensive field evaluation demonstrates