| description abstract | The basic objective of long-term optimization of large-scale hydropower system operations (LOLHSO) is to increase water resource use efficiency and construct a clean and low-carbon-intensity energy system. However, limited by rather stochastic inflows, large-scale hydropower systems, and complex objectives and constraints, LOLHSO is characterized by uncertainty, high dimensionality, nonlinearity, and nonconvexity, which pose great challenges in modeling. This paper presents an improved hybrid decomposition-coordination and discrete differential dynamic programming model (IDC-DDDP) for solving the LOLHSO problem. In IDC-DDDP, a decomposition strategy considering the adjustment potential of hydropower plants is designed to reduce the system size. Meanwhile, owing to the sensitivity of DDDP to initial trajectories, a data mining–based strategy is developed as a means of generating superior initial trajectories. A corridor generation strategy is presented to determine the discrete steps and enhance global search abilities. Case studies in a hydropower system in southwestwern China demonstrate the practicability and robustness of the proposed model. | |
