Water-state-aware coordinated scheduling of cascade hydropower and pumped storage for power systems with high renewable penetration

This is the authors' abstract. We don't add key points for this paper.

High shares of variable renewable energy (VRE) increase flexibility requirements and challenge the coordinated operation of cascade hydropower and pumped storage (CH-PS). This study proposes a multi-timescale coordinated scheduling framework for CH-PS systems to improve renewable integration while preserving intertemporal water value and supporting short-term power balancing. The framework links a long-term water allocation layer with a short-term dispatch layer through a water-state-aware bargaining mechanism and a dynamic water-value signal, enabling storage conditions and system tightness to be translated into coordinated dispatch targets. To solve the short-term constrained scheduling problem, a Lagrangian reinforcement learning method with PID-adaptive dual updates is developed, together with rolling-horizon forward evaluation for intraday correction. A case study based on a representative system in Fujian Province, China, shows that the proposed strategy achieves full VRE accommodation, reduces the share of purchased electricity in total load from 30.12% to 12.16%, reduces the residual-load standard deviation by 66.62%, and increases annual net revenue by 29.25%. These results are based on historical data from this regional system and indicate that coordinated CH-PS operation can enhance system flexibility and renewable integration while improving economic performance; further validation against optimization baselines and forecast-error scenarios is required.