Akademik Veri Yönetim Sistemi
9th International Conference on Mathematical Advances and Applications (ICOMAA-2026), İstanbul, Türkiye, 6 - 08 Mayıs 2026, cilt.9, ss.1-10, (Tam Metin Bildiri)
This study presents a stochastic multi-objective optimization and predictive control framework for the integration of second-life batteries (SLBs) into grid-connected photovoltaic (PV)-based workplace electric vehicle (EV) charging systems. The main methodological challenge addressed in this work is the heterogeneous state-of-health (SoH) behavior of retired battery modules, which is commonly simplified or ignored in conventional homogeneous energy storage models. To represent this uncertainty, a stochastic module-level SoH modeling approach is developed by considering variations in available capacity and internal resistance. Based on this representation, a multi-objective optimization problem is formulated to simultaneously minimize net present cost and carbon emissions while maximizing the self-sufficiency ratio. The resulting optimization problem is solved using the NSGA-III algorithm. In addition, a model predictive control (MPC) strategy with a 24-hour prediction horizon is implemented to improve the operational decision-making of the PV–SLB–EV charging system under dynamic solar generation and stochastic EV demand conditions. Simulation results based on annual meteorological data and probabilistic EV charging profiles show that the proposed framework enhances system efficiency, reduces grid dependency and emissions, and supports more reliable utilization of second-life batteries compared with conventional homogeneous battery modeling approaches.