Abstract:
This study proposes a design flood estimation method for cascade reservoirs during operation period to maximize the comprehensive benefits both for flood control and hydropower generation.The joint probability distributions of floods occurring at all sub-basins are established using the multivariate t-Copula function. The most likely regional composition is derived using the Monte Carlo simulation and the Genetic Algorithm (GA) methods. The design floods as well as the flood limited water levels of the Wudongde-Baihetan-Xiluodu-Xiangjiaba cascade reservoirs duringoperation period are analyzed. It is shown that:① The results of the most likely flood regional composition are reasonable and can offer a meaningful reference for the analysis of design flood estimation for cascade reservoirs.② The design floods of downstream reservoirs are significantly influenced by the operation of upstream reservoirs. The 1000-year design flood peak,3 d,7 d and 30 d design flood volumes at Xiangjiaba reservoir significantly decrease by 15 400 m
3/s (35%),3.56 billon m
3(33%),7.02 billon m
3(30%) and 8.50 billion m
3 (11%),respectively.③ The obtained flood control water levels of downstream reservoirs are slightly higher than flood limited water levels but without reducing the original flood prevention standards. The flood control water levels (flood limited water levels) of Baihetan,Xiluodu and Xiangjiaba reservoirs during operation period are 788.82 m(785 m)、570.67 m(560 m) and 371.36 m(370 m),which can increase the hydropower generation by 2.1%,6.1% and 1.4%,respectively,and in total 1.71 billion kW·h annually.