Optimization of electric vehicle charging stations in terms of location and charging time

The integration of electric vehicles (EVs) into power systems introduces new load patterns, creating significant challenges for the planning and operation of distribution networks. This thesis investigates the technical impacts of Electric Vehicle Charging Stations (EVCS) on distribution systems and proposes an optimization-based approach to minimize total active power losses through spatial and temporal planning. The study is conducted using the IEEE 33-bus test system as the modeling framework. In the first phase, five different scenarios were developed by randomly placing EVCS units on the network, beginning with five charging stations and reducing the number progressively down to one. For each case, total power losses were calculated. It was observed that random placement leads to substantial increases in system losses. To address this, the Particle Swarm Optimization (PSO) algorithm was applied to determine the optimal EVCS locations, resulting in significantly lower losses. These PSO results were compared with those obtained using the Genetic Algorithm (GA), and it was found that both algorithms identified the same optimal bus locations, validating the robustness of the proposed approach. In the second phase, a time-based charging scenario was introduced. Assuming each vehicle charges for 8 hours within a 24-hour period, the GA was used to determine the most suitable charging intervals for each EVCS. During this analysis, the EVCS locations obtained from the PSO algorithm were kept fixed. Furthermore, a photovoltaic (PV) generation unit was integrated at each charging station to alleviate the load on the grid and support renewable energy utilization. The optimal charging schedules derived by GA were then compared with fixed 8-hour charging profiles, demonstrating that the optimized schedules led to significantly lower power losses. The results of this study highlight that simultaneous optimization of both spatial placement and temporal operation of EVCS units can lead to considerable improvements in power loss reduction and voltage profile regulation within distribution systems. This thesis demonstrates that heuristic algorithms such as PSO and GA can be effectively utilized for intelligent planning of EV charging infrastructure and offers a practical foundation for future smart grid applications.