Optimal vehicle fleet and charging strategy for electric ground service equipment at airports based on flight schedule

As the aviation industry transitions towards a green and low-carbon future, deploying electric ground service equipment (E-GSE) and integrating photovoltaic (PV) generation at airports can reduce energy consumption and emissions. This study proposes a two-stage optimization framework. In the first stage, a long short-term memory (LSTM) neural network is employed to forecast seasonal PV power generation, aiming to mitigate the impact of PV output volatility on downstream decision-making. In the second stage, a techno-economic optimization model is formulated, incorporating vehicle operation time constraints, time-of-use electricity pricing, and PV generation, to minimize the annual total cost for airport ground operation. The framework simultaneously optimizes vehicle procurement, charger allocation, and charging schedules. To enhance computational efficiency, the model is linearized, and a Benders decomposition-based mixed-integer programming approach is introduced to solve large-scale instances efficiently. A case study based on real flight schedule data from Guangzhou Baiyun International Airport demonstrates the effectiveness of the proposed method. The proposed framework identifies an optimal airside PV capacity of 559.60 kW, reducing the airport ground operator's annual total cost by 624,042.60 CNY relative to the no-PV case and achieving a PV payback period of 3.04 years. Compared with conventional vehicle-to-pile ratio planning, the optimized strategy reduces the E-GSE fleet from 198 to 162 and decreases the number of fast chargers from 89 to 73, corresponding to reductions of approximately 18% in both categories.