The rapid growth of electric vehicles (EVs) demands intelligent, efficient, and reliable charging... more The rapid growth of electric vehicles (EVs) demands intelligent, efficient, and reliable charging infrastructure. Conventional EV charging systems suffer from high switching losses, grid instability, thermal stress, and poor energy management. This paper proposes an AI-based adaptive power electronics control system for smart EV fast charging stations. The proposed system integrates a multi-layer neural network controller with an interleaved bidirectional DC-DC converter using Silicon Carbide (SiC) MOSFETs and a three-phase grid-connected inverter. The AI controller optimizes switching frequency, charging current, thermal management, and reactive power compensation in real time. Simulation results in MATLAB/ Simulink demonstrate 96.4% peak converter efficiency, Total Harmonic Distortion (THD) below 3.5% (compliant with IEEE Std 519-2022), a 26.9% reduction in charging time, 18°C maximum battery temperature rise, and a 44.7% reduction in converter power losses compared to conventional PI controllers. The system supports bidirectional energy flow for Vehicle-to-Grid (V2G) operation and seamless renewable energy integration, contributing toward future intelligent EV ecosystems for smart cities.
This paper presents a Multi-Level Priority-Based Smart Energy Management System (SEMS) for reside... more This paper presents a Multi-Level Priority-Based Smart Energy Management System (SEMS) for residential solar photovoltaic (PV) systems integrated with battery storage. The proposed system classifies residential loads into critical and non-critical categories and employs a rulebased control algorithm to manage power distribution based on real-time solar generation, battery state of charge (SOC), and load demand. Critical loads such as medical devices, refrigerators, and essential lighting receive uninterrupted power supply under all operating conditions, while non-critical loads are managed according to energy availability. The system also incorporates mechanisms to prevent the voltage instability caused by simultaneous switching of high-power appliances, addressing the crowding phenomenon in smart grids. Simulation results demonstrate that the proposed SEMS significantly improves energy efficiency, ensures reliable power delivery to critical loads, and optimizes battery utilization compared to conventional systems without energy management. The simple rule-based design makes the system well-suited for practical residential deployment without requiring complex computational infrastructure.
Uploads
Papers by Raunak Tiwari