| One barrier to the optimal operation of hydropower plants is a lack of accurate inflow forecast information. This is true for both seasonal inflow expectations, which affects long-term planning for bulk energy production, and daily inflow expectations, which affects flexibility, and is further exacerbated in the case of cascading plants. The proposed work aims to develop: 1) accurate machine-learning based closed-loop forecast models for seasonal and day-ahead water inflows; and 2) enhanced cascading hydroelectric system (CHE) modeling and data-driven optimization approaches to explore multidimensional spatial-temporal hydropower operational flexibility potentials with proper consideration of unique characteristics of CHE systems. |
