论文部分内容阅读
区域洪水频率分析能够提高实测资料过短的站点或无资料地区的设计洪水的代表性。文章选取了长江上游川江段13个水文站的最大洪峰流量系列,应用模糊c均值(FCM,FuzZY c-Means)类型算法进行水文相似性分区,选用洪水系列的L-变差系数t_2和L-偏态系数t_3作为水文站点分区的模糊聚类因子,并利用线性矩法估计洪水频率曲线的参数,推求长江上游川江段的区域洪水频率曲线和设计洪水。
Regional flood frequency analysis can improve the representativeness of design floods at sites where measured data are too short or in regions without data. In this paper, the series of maximal peak discharges of 13 hydrological stations in the upper reaches of the Yangtze River are selected. Hydrological similarity partitioning is carried out by the FCM (FuzzZy c-Means) type algorithm. The L-variation coefficient t 2 and L- The skew coefficient t_3 is used as the fuzzy clustering factor of hydrological station zoning, and the parameters of the flood frequency curve are estimated by using the linear moment method to infer the regional flood frequency curve and designed flood in the upper reaches of the Yangtze River.