Probabilistic energy management of renewable-based charging stations hub for urban electric public transportation fleets

Urban public transport electrification increasingly couples charging demand with variable renewable supply, creating operating conditions that deterministic station scheduling does not represent well. This work addresses that limitation through a probabilistic energy-management framework for a grid-connected charging station supplied by photovoltaic (PV) and wind generation. Solar irradiance and wind speed were represented by Beta and Weibull distributions, while fleet-side uncertainty was modeled through the initial state of charge and price-responsive charging or discharging behavior. Monte Carlo simulation generated stochastic realizations, and the Coati Optimization Algorithm (COA) determined 24-hour grid exchange and fleet power decisions subject to power-balance and operating constraints. The same scheduling problem was solved under deterministic and uncertainty-aware scenarios and benchmarked against four metaheuristic methods. Uncertainty raised the COA operating cost from 190.97 USD to 197.84 USD, indicating the economic effect of simultaneous supply- and demand-side variability. Even so, COA retained the lowest cost among the tested methods; under uncertainty, its cost was 15.70% below that obtained by particle swarm optimization. The stochastic results also preserved hourly power balance while capturing changes in renewable output and bidirectional fleet demand. The framework therefore provides a tractable basis for uncertainty-aware scheduling of renewable-supported public-transport charging stations and for assessing optimizer behavior under coupled generation and charging variability.