Design and optimization of a renewable energy based electric vehicle fast charging station

Abstract

This research addresses the optimization challenge of renewable energy-based electric vehicle fast charging stations (EVFCS) using a multi-objective approach. The core contribution lies in the development of two Enhanced Multi-Objective Particle Swarm Optimization (EMOPSO) algorithm, specifically tailored to harmonize conflicting objectives encompassing cost efficiency, reduction of greenhouse gas emissions, and maximal exploitation of renewable energy resources. The research unfolds in distinct phases. Firstly, it proposes an innovative design for EVFCS, blending wind and photovoltaic power generation with BESS, aligning renewable energies with energy storage. The EV charging load simulation model adapts to time-of-use electricity prices, reflecting regulatory dynamics. The integration of a Battery Management System (BMS) further optimizes BESS performance. In most iterations, both versions of EMOPSO produce nearly identical solutions with minimal deviations. In this specific instance, the optimal solution recommends quantities of 10 for wind turbines, 30 for photovoltaic (PV) modules, and approximately 17 for batteries. Under this configuration, the achieved objective values are as follows: Cost of Energy (COE) is 9.50 bdt/kWh, Greenhouse Gas Emissions (GHG) is 0.076 kg/kWh, total space required for wind turbine and solar PV installation is 3020 m², and maximum grid electricity demand is 120.98 kW. The power flow analysis reveals that the Electric Vehicle Fast Charging Station (EVFCS) has the capacity to both import and export power from the grid. Employing the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), a systematic scenario analysis was conducted, and the proposed model secured a TOPSIS rank of 1 when comparing grid-only charging, off-grid solutions, and the absence of time-of-use pricing dynamics. The sensitivity analysis indicates that both versions of EMOPSO are responsive to investor preferences, and by increasing the decision weight of any objective, it is possible to further minimize that cost function. A comparative analysis among EMOPSO version 1, version 2, traditional Multi-Objective Particle Swarm Optimization, and Genetic Algorithm was performed. The Diversity of Solutions (D) is as follows: 1144, 656, 1277, and 3136, respectively. The Spacing Metric (S) is as follows: 142, 112, 280, and 1098, respectively. Lower values of D and S indicate more favorable outcomes. It is unequivocal that the EMOPSO methods yield the most compact values for both D and S. This result emphasizes the compatibility of the EMOPSO approaches with the model developed in this study.