Techno-economic optimization of PHEV parking lot siting and sizing in smart microgrids using cuckoo search algorithm

In recent years, electric vehicles have become more prevalent in the transportation fleets of the world's developed nations. When incorporated into microgrids, they can maximize productivity, and provide additional flexibility when optimally sited. However, poor placement and sizing of the EV parking lots will often, as was proven in this paper, increase microgrid losses, decrease power quality, and increase the overall cost of operations. This paper sets out to optimize the placement and capacity of plug-in hybrid electric vehicle (PHEV) parking at an IEEE 33-bus microgrid using cuckoo search (CSA) and genetic algorithms (GA) in two separate scenarios; with no option of selling energy to the mother grid and in a scenario where it is allowed to export energy. The results showed that optimal placements of EV parking optimize microgrid costs and reduces losses. It was consistently seen that the CSA showed better performance than the GA for both scenarios, in obtaining the optimal cost more efficiently than the GA. One of the main contributions of this study is the application of CSA for the techno-economics adaptation of PHEV parking lots, considering operational/technical and economic constraints and showing better performance than conventional optimization methods. • The study proposes the use of Cuckoo Search Algorithm (CSA) for optimal siting and sizing of PHEV (plug-in hybrid electric vehicle) parking lots in smart microgrids. • CSA outperforms other metaheuristic algorithms (GA, PSO, DE, SA) in reducing total cost, energy losses, and CPU execution time. • Optimized PHEV parking lots improve voltage stability, balance active and reactive power, and enhance microgrid operational efficiency. • Scenario 2 demonstrates a 16 % reduction in total cost and a 24 % reduction in energy losses, highlighting both economic and technical benefits. • Scalability and robustness of CSA validated on microgrids up to 300 buses, showing stable convergence and computational efficiency.