Seite - 267 - in Book of Full Papers - Symposium Hydro Engineering

3. CASE STUDY Based on the abundant water power resources and mountainous topography in middle reach of the Jinsha River, China, four concrete gravity dams(A, B, C, D) were built for the development and utilization of hydropower resources. The four dams lie among three hydrological stations: Shigu, Jinjiangjie and Panzhihua, making a series of cascade reservoirs. According to feasibility studies, the four dams are all scale first-grade first-class projects with check standard for flood control equals to 5000-year return period. 3.1. PARAMETERS OF REPRESENTATIVE HYDROLOGICAL STATION Annual flood peaks recorded in the 3 representative hydrological stations during 1957 and 2005 [11] are gained by P-III distribution curve fitting which deduces the best-fitting EX, Cv and Cs for each hydrological station. Results are shown in Table 1. Table 1 Parameters of curve fitting Hydrological station Parameters Flood peak in different frequency(m3·s-1) EX Cv Cs/Cv 0.01% 0.02% 0.2% Shigu 5140 0.29 4 14600 13900 11500 Jinjiangjie 6710 0.31 4 20200 19200 15800 Panzhihua 7360 0.31 4 22200 21100 17300 3.2. STOCHASTIC SIMULATIONS FOR REPRESENTATIVE HYDROLOGICAL STATION 3.2.1 Climate change In consideration of climate change, GAMLSS[12,13] is used to identify the optimum flood distribution[14], reveals whether time series is stationary or not. Firstly, three probability distribution functions (PDF) are selected, namely: (1)Gumbel distribution (2)two parameter LogNormal(LN2) (3)two parameter Pearson Type II (P2), Gamma distribution Secondly, time trend is assumed in the first two moments (the mean value and the standard deviation): (1) trends in the mean value (A) (2) trends in the standard deviation (B) 267