基于模糊时序有效聚类的太湖流域汛期分期研究

Novel flood season division method based on fuzzy time series-effective cluster and its application to Taihu lake basin

依据流域暴雨洪水的长期规律对汛期进行合理划分,是确定分期设计洪水与分期汛限水位,进而制定洪水资源利用方案,挖掘洪水资源利用潜力的前提。本文通过引入时间维聚类算子对模糊C-均值聚类方法,以及有效聚类判别函数进行改进,并为消除模糊C-均值聚类方法对初值敏感的问题,用实码加速遗传算法替代迭代方法进行求解,在此基础上建立了基于模糊时序有效聚类的汛期分期方法。将该方法应用于太湖流域汛期分期中,为消除汛期起始点对分期结果敏感性和人为干扰,应用时以防洪规划规定的汛期起始点5月1日为中点,前后推移不同时段构成11个方案分别进行研究。结果表明:本方法有效,提出太湖流域汛期划分为5期的计算结果,经分析是合理的。

Seasonal distribution of precipitation in eastern China affected by monsoon climate is uneven. Formulation of operation scheme of flood level control according to seasonal flood variations is important to flood resources utilization aiming at easing water shortages and improving water environment. Division of basin flood season, a muhi-dimensional time series clustering problem with unknown number of cluster, is the premise of designing stage flood and stage flood control level of large water project. Thus, cluster methods are suitable to handle multi-dimensionality, time series and clustering validity, but no one meets all these needs currently so far. This paper puts forward a dynamic fuzzy C-means clustering method and clustering validity function-based time series, based on Xie＇s previous studies. There remains, however, a complex nonlinear optimal problem in this dynamic method, making it very difficult to solve with regular optimal method. To solve such problem, this paper adopts a real coding-based accelerating genetic algorithm, a very effective algorithm that can overcome the difficulty of initial value sensitivity in regular iterative method for optimization of fuzzy C-means clustering. Thus, by coupling all the methods and techniques above, we establish a new method for flood division, namely dividing flood season method based on fuzzy time series-effective cluster. We also present a set of self-contained steps of model development and an example application to the Taihu basin. This application shows that the new method is very effective and that the dividing scheme so obtained is reasonable.