Two-stage data-driven distributionally robust optimization for capacity configuration of pumped-storage used to retrofit cascade hydropower stations

The increasing penetration of wind and photovoltaic power poses significant challenges to power system operation due to their inherent uncertainty and variability. Retrofitting cascade hydropower stations with pumped-storage units offers an effective way to enhance system flexibility; however, the optimal sizing of pumped-storage units under renewable energy uncertainty remains insufficiently explored. This paper proposes a two-stage data-driven distributionally robust optimization framework for the capacity configuration of pumped-storage units in a hybrid wind-photovoltaic-hydro-thermal system. The proposed framework jointly optimizes capacity investment and operational decisions while characterizing renewable energy uncertainty through a data-driven ambiguity set constructed from historical data. Case studies show that the proposed approach significantly improves both economic performance and operational flexibility. Compared with the benchmark case, the total system cost is reduced by 41.12%. The results demonstrate that the proposed framework provides an effective and robust planning tool for flexible resource deployment in power systems with high renewable energy penetration.