Real-time dispatch of cascade hydropower-photovoltaic complementary system with hybrid pumped-storage based on Wasserstein distributionally robust optimization

The global transition to renewable energy highlights the critical need to integrate variable photovoltaic (PV) generation with flexible cascade hydropower (CHP) systems. While CHP systems offer operational advantages for PV complementarity, increasing PV penetration further challenges their flexibility. Hybrid pumped-storage (HPS) retrofitted within existing CHP systems enhances bidirectional regulation and adjustment capability. However, hedging against PV power uncertainties during real-time operation is a significant challenge with consideration of spatiotemporal electrical and hydraulic coupling. This paper proposes a real-time dispatch approach for CHP-PV-HPS complementary systems based on Wasserstein distributionally robust optimization (WDRO) with a model predictive control framework (MPCF). WDRO constructs data-driven ambiguity sets to balance robustness and optimality, while MPCF updates dispatch decisions using rolling-horizon optimization. To ensure hydraulic constraint feasibility, a feasibility violation detector (FVD) algorithm is proposed and the nonconvex model is reformulated into a mixed-integer linear programming problem to ensure computational tractability. Simulation studies in a real system in southwest China are conducted to produce online dispatch policies. The results verify the effectiveness of HPS in reducing deviation and volatility of integrated power output. Besides, the proposed approach can effectively prevent units power from entering forbidden zones and reservoir water level from violating boundary constraints, while maintaining stable performance across both in-sample and out-of-sample tests. This work advances the operational safety and efficiency of renewable-dominated complementary systems integrated into grids, offering an adaptive solution for real-time energy management under uncertainty.