Sustainable Solar-Powered EV Charging System Design Using Machine Learning, DC Fast Charging, and an Intelligent DMPPT Optimization Technique

The increasing demand for electric vehicles (EVs) necessitates the development of efficient and sustainable charging infrastructure. This dissertation presents a comprehensive approach to designing a solar-powered EV fast charging station. This dissertation proposes novel techniques, including a hybrid approach combining the Random Forest Regression (RFR) and Long Short-Term Memory (LSTM) models for accurate prediction of solar photovoltaic (PV) power generation, an intelligent maximum power point tracking (MPPT) to increase system efficiency, and a comprehensive design and analysis of the solar-powered EV charging station system. Accurate solar PV power generation prediction is crucial for efficient energy management and grid integration. This dissertation proposes a hybrid approach combining the RFR and LSTM models, which accurately estimate solar power output by capturing complex relationships and patterns in solar power generation data. The hybrid model overcomes the limitations of individual techniques and achieves enhanced accuracy in long-term solar power forecasting. Real-world data evaluation demonstrates its superior performance compared to standalone models, highlighting its potential for supporting renewable energy planning and management. This dissertation also introduces an intelligent MPPT technique that integrates fuzzy optimization with methods such as Artificial Bee Colony (ABC) and Genetic Algorithm (GA). This technique enhances MPPT efficiency and accuracy, maximizing energy extraction from PV systems operating under dynamic conditions and partial shading. Simulation results validate the technique's superior performance, reducing power losses and improving tracking accuracy. Furthermore, a comprehensive design and analysis of the solar-powered DC fast EV charging station system is presented. The design integrates advanced power conversion techniques, including a forward converter, three-phase inverter, three-phase rectifier, and bidirectional half-bridge DC-DC converter. These components enable efficient power management, high power density, bidirectional power flow, and suitable integration with the power grid. The proposed system offers enhanced energy utilization, reliable grid integration, and sustainable EV charging infrastructure, contributing to the reduction of carbon emissions. By combining these three studies, this dissertation provides valuable insights into the design and optimization of solar-powered EV charging stations. The proposed methodologies contribute to advancing renewable energy utilization, accurate solar PV power generation prediction, and developing efficient and sustainable EV charging infrastructure. These efforts aim to accelerate the adoption of EVs and promote a greener future.