Abstract: For a river-type reservoir, a reasonable hydropower generation scheduling should consider the effects of its dynamic reservoir capacity and unsteady flow propagation in the mainstream and tributaries within a certain area. Aimed at such an issue, this paper constructs a reservoir optimal power generation scheduling model coupled with a one-dimensional river network unsteady flow model, which is solved using an OpenMP-based parallel dynamic programming algorithm and applied to a case study of the Three Gorges Reservoir (TGR). The results show this dynamic capacity model is effective in simulations of the spatial-temporal variations of discharges and water levels along the mainstream and tributaries within the TGR area. Compared with the static capacity model, it shows distinct differences in power generation process across operational periods—more stable variations in drawdown period and flood period, while experiencing significantly more fluctuations during impoundment period. Specifically, in these three periods, the total power generation decreases by 206 million kW·h, 129 million kW·h, and 47 million kW·h, respectively, with the most significant impact observed during drawdown period. And, during the initial stages of drawdown and impoundment, the model enhances power generation due to an increase in outflow discharges. This study would help refined simulations of the hydropower generation scheduling for river-type reservoirs.

Key words: power generation optimization, river-type reservoirs, dynamic capacity, one-dimensional unsteady flow in river networks, parallel dynamic programming