Short-term net load smoothing via stochastic scheduling of runoff-contained cascade hydropower under wind-photovoltaic uncertainty

Cascade hydropower provides critical flexible resource to mitigate grid load fluctuations caused by the large-scale integration of wind and photovoltaic power. This paper proposes a scenario-based stochastic programming model for the short-term peak shaving problem of cascade hydropower with run-of-river stations under wind-PV uncertainty. First, Latin hypercube sampling and synchronous backward reduction algorithm are employed to generate and condense wind-photovoltaic power scenarios characterizing uncertainty. Second, a stochastic optimization model is formulated to minimize the expected peak-to-valley difference in residual load, where nonlinear and nonconvex constraints are linearized using multiple linearization techniques, transforming the model into a mixed integer linear programming problem. Finally, simulations on a real-world cascade hydropower system in China demonstrate that the proposed model effectively handles uncertainty and smooths net load across various scenarios, achieving an average peak-to-valley difference reduction of 37.94%. Compared with a deterministic optimization model, the stochastic programming model achieves a peak-to-valley difference reduction rate of 14.63% in the simulation experiments, demonstrating a stable load smoothing effect. Moreover, experiments reveal that the limited regulating capacity of run-of-river stations contributes positively to the overall peak shaving performance of the cascade system.