Performance analysis of a PV powered drone delivery network under realistic solar energy constraints

As the global shipping industry aims to reduce greenhouse gas emissions by 50 % by 2050, drone delivery networks present a promising solution for last-mile logistics. However, existing studies often rely on idealized assumptions regarding energy availability, overlooking the stochastic nature of renewable energy sources. This paper presents a comprehensive performance analysis of a PV-powered drone delivery network in the North Adriatic Sea, utilizing a discrete-time, agent-based simulation model. The study integrates real-world, time-series photovoltaic generation data to evaluate system throughput, energy consumption, and charging bottlenecks over a 48-hour period. Results indicate a severe performance gap between idealized and realistic scenarios: while an ideal system achieved a 100 % success rate, the introduction of realistic solar variability reduced the success rate to 51.7 % and caused significant queuing delays totaling over 223 h. Furthermore, the study reveals a counter-intuitive finding where increasing the number of chargers per hub during low-energy periods further degraded system performance by distributing scarce energy too thinly. A scalability analysis across fleet sizes of 30 to 240 drones demonstrates that proportional PV scaling restores performance at moderate fleet sizes, but a hub topology ceiling emerges at larger scales, indicating that routing and congestion management become the binding constraints beyond a critical fleet density. These findings underscore the critical necessity of incorporating energy intermittency constraints into the design of sustainable autonomous logistics infrastructure and identify techno-economic optimization of PV sizing and hub topology as key directions for future work.