改进的模糊模式识别在金沙江径流特性分析中的应用

Application of Improved Fuzzy Pattern Recognition to Runoff Characteristic Analysis of the Jinsha River

以模糊c均值算法为基础,通过改进隶属度函数的方法,将金沙江小黄瓜园45年的月平均径流资料进行多维空间中聚类。在聚类基础上,采用基于模糊模式识别的分析方法,得出了该站径流系列的多维丰枯特性及其变化周期。其计算结果与实测径流资料对比,较为符合实际情况,表明该方法能够有效分析径流的周期特性。

Multidimensional cluster was made for the 45 years＇ mean monthly runoff data acquired at Xiaohuangguayuan hydrological station on the Jinsha river by improving the membership function on the basis of the fuzzy c averaging method. The paper uses the fuzzy pattern recognitionbased analysis method to determine the multidimensional high and low water characteristics and their variation period of the runoff series at that hydrological station based on the cluster. Compared with the measured runoff data, the calculation results conform to reality, showing that the method can be used to make an effective analysis of runoff＇ s periodic characteristic.