聚类识别算法解读洪水峰型

Analysis of the Flood Peak Type by Clustering Algorithms

  • 摘要: 洪水峰型的解读是项重要的基础工作。本文提出洪水峰型指标的界定原则,并从多角度提取反映洪水峰型的指标。采用人工鱼群优化的投影寻踪模型解析洪水峰型,结果说明:采用分类的研究思想,仅7项洪水峰型指标即可辨识出武江流域1955-2007年的53场洪水中,有单峰型洪水28场、双峰型洪水21场和三峰型洪水4场。在此基础上,利用辨析双峰型差异的4项指标,进一步识别21场双峰型洪水的峰型差异,得出I型双峰型洪水11场,II型双峰型洪水10场。

     

    Abstract: The flood peak types provided the premise and foundation for the reservoir operation rules. This paper explained the method of selecting the index of flood peak type. According to the 7 indicators,this paper identifieed the flood peak types by Projection Pursuit based on Artificial Fish Swarm Algorithm. In the Wujiang River Basin,the simply peak type,double peaks type and three peaks type is 28 floods,21 floods and 4 floods,respectively. To further analysis the differences between the double peaks type,four indicators,which differentiated the double peaks type,were used to identified that 11 floods belong to I-double-peaks type and the rest 10 floods belong to II-double-peaks type.

     

/

返回文章
返回