Chance-constrained receding-horizon medium-term cascaded hydropower scheduling

Scheduling cascade hydropower systems faces significant challenges arising from complex hydraulic coupling and the increasing volatility of water inflows driven by climate variability. This paper presents a receding-horizon optimization framework for medium-term scheduling of cascade hydropower systems under temporally dependent inflow uncertainty. Traditional open-loop multi-time-scale scheduling neglects week-to-week variations in inflow distributions, which leads to biased probabilistic assessments and suboptimal operational decisions. To address this issue, a conditional density estimation method is developed to characterize the evolution of inflow uncertainty and incorporate it into a receding-horizon scheduling model. A Probabilistic Control Barrier Function (PCBF) reformulation is introduced to ensure probabilistic recursive feasibility, enabling safe operation as the scheduling horizon recedes. To obtain a tractable implementation, a Conditional Sample Average Approximation (CSAA) of the PCBF-based problem is constructed, and the uniform convergence of its optimal solutions to those of the ideal formulation is established. Real-data case studies on cascade hydropower systems demonstrate that the proposed method effectively adapts to evolving hydrological conditions and improves the reliability of medium-term scheduling under uncertainty. • Receding-horizon medium-term scheduling for cascade hydropower systems. • Conditional density modeling of temporally dependent water inflows. • PCBF-based formulation ensuring probabilistic recursive feasibility. • CSAA-based approximation with provable uniform convergence. • Validation using real-world cascade hydropower inflow data.