Research On Uncertainty Of Reservoir Flood Control Risk In Mountain Area Of Northern Slope Of Tianshan Mountain Under Climate Change

The change of climate affects the confluence process of Manas River,which makes the hydrological series change and the consistency hypothesis is no longer satisfied.In this paper,based on the hydrometeorological data of 1956 ~ 2014 a in the Kenswatt control section of the Manas River,the statistical characteristics are analyzed,and the variation diagnosis of the annual maximum flood volume sequence is carried out in combination with the hydrophysical mechanism.The "decomposition-synthesis" theory is used to modify the inconsistent annual maximum flood volume sequence.The annual maximum flood volume sequence under the present condition is fitted with the traditional P-? distribution by Markov Chain Monte Carlo(MCMC)method based on Bayesian theory,and the uncertainty of the parameters is analyzed.The design values of the flood volume under different recurrence periods are obtained.According to the flood regulation of the reservoir,the flood diversion is carried out and the risk analysis is carried out.The specific contents and results are as follows:(1)Consistency check of hydrological series.By means of variable fuzzy set test and Pettitt nonparametric test,it is found that there is a jump variation in the annual maximum flood volume sequence of Manas River,in which the annual maximum 1st flood volume sequence varied in 1993.The variation of the annual maximum flood volume sequence on the 3rd,7th and 15 th of the year occurred in 1995.Combined with Mann-Kendall test,it is considered that the variation trend of the sub-sequence after the variation point of the annual maximum flood volume sequence is not significant,combined with the hydrophysical mechanism,and climate change is the main reason for the variation of annual maximum flood volume series.(2)Uncertainty analysis of P-? distribution parameters.The "decomposition-synthesis" theory is used to modify the inconsistent annual maximum flood volume sequence,and the annual maximum flood volume sequence under the present condition is obtained.The Markov Chain Monte Carlo method based on Bayesian theory is used to fit the modified sequence with P-? distribution,in which Gibbs-MCMC and AM-MCMC methods are used for sampling,and the performance of the two sampling methods is compared by three indexes.Then the uncertainty of quantitative parameters of mean and 95% confidence interval is given.(3)Limit flood control risk analysis of Kenswatt reservoir.Taking a flood in 1996 as a typical flood process,the design flood at different frequencies is obtained by the same frequency amplification method,which is regarded as the reservoir inflow flood process,and the flood routing calculation is carried out according to the reservoir flood control operation rules.The frequency analysis method is used to analyze the ultimate flood control risk rate of the reservoir under the present conditions.the ultimate flood control risk rate is 0.1237% under the present conditions.at the same time,the fuzzy interval of the ultimate flood control risk is given by using the triangular fuzzy number risk analysis method.