Flood control ability of river-type reservoirs using stochastic flood simulation and dynamic capacity flood regulation

Abstract

The flood control ability of river-type reservoirs is greatly influenced by the inflow flood characteristics and the calculation of the dynamic capacity flood regulation. The current research on the flood control ability of large river-type reservoirs mainly uses a design inflow flood based on the enlargement of typical historical floods. This approach does not adequately consider the different forms of inflow floods and the flood regulation of dynamic capacity. To address this issue, the Three Gorges Reservoir (TGR) was used as a case study. Measured daily runoff data over a 60-year period from the Zhutuo, Beibei, and Wulong Hydrological Stations were used. A stochastic model of inflow floods for the TGR was established using the periodic stationary autoregressive method, which generated 4,000 stochastic inflow floods. A dynamic capacity flood regulation capacity model was used to simulate different inflow floods (the stochastic floods and the measured flood for 1954. The results were based on the 1,000-year flood, which was enlarged by the stochastic floods. The results show that, under the most dangerous inflow flood conditions, the maximum discharge at the Zhicheng Hydrological Station was 75,550 m3/s, which is 7,850 m3/s greater than the maximum discharge at Zhicheng for the 1954 flood. The results also showed that the probability of a more dangerous inflow flood than the 1954 flood occurring is 6.63%. This shows that it is not sufficient to calculate the flood control ability of river-type reservoirs using only historical flood measurements. In flood control design, it is necessary to use stochastic flood simulation methods to find the most dangerous inflow flood process of a reservoir at a specific design frequency, and to simulate the propagation of the flood in the reservoir through dynamic capacity flood regulation. These methods enable the flood control ability of large river-type reservoirs to be demonstrated.