基于EMD的HHT变换技术在长江三峡水库年平均流量预报中的应用  

Application of EMD-based Hilbert-Huang Transform Technology in Forecasting Mean Annual Discharge of the Three Gorges Reservoir on the Yangtze River

作　　者：曹丽青[1,2] 林振山[1] 

机构地区：[1]南京师范大学地理科学学院,江苏南京210046 [2]河海大学水文与水资源学院,江苏南京210098

出　　处：《水文》2008年第6期21-24,共4页Journal of China Hydrology

基　　金：高等学校博士学科点专项科研基金资助课题(20060139010)

摘　　要：本文应用基于经验模态分解(EMD)的Hilbert—Huang变换(HHT)技术,对长江宜昌站1900～2000年来天然年平均流量序列进行了分析研究,并在此基础上建立了长江三峡水库年平均流量多尺度统计动力预报模型。研究结果表明:(1)长江三峡年平均流量存在准3.2年、7.4年、13.1年、16.7和53.8年的周期,并以准3.2年和7.4年的波动为主;(2)基于EMD的HHT变换技术的三峡水库年平均流量多尺度统计动力预报模型,预报效果较好,对2001～2004年进行试验预报,预报模型的预报相对误差在2.4%以下。本文研究成果对长江三峡水库防洪调度以及发电计划制作有重要的参考价值。A novel multi-scale analysis method, Hilbter-Huang Transform （HHT） based on EMD, was used to analyze the mean annual discharge of the Three Gorges Reservoir on the Yangtze River in China. The multi - scale statistic and dynamic model for mean annual discharge was established based on HHT method. The results show that （1） The mean annual discharge of the Three Gorges reservoir can be decomposed into four timescales quasi-periodic oscillations including a 3.2 year signal, a 7.4 year signal, a 13.1 year signal, a 16.7 year signal and a 53.8 year signal, as well as a trend. With each contributing ration of the quasiperiodicity discussed, timescale oscillation of the 3.2 year and the 7.4 year is the most prominent; （2） The multiscale statistic and dynamic model for mean annual discharge of the Three Gorges Reservoir based on HHT was checked by observed data from 2001 to 2004. The results show that the model and method are effective in practice. The relative errors of model are lower than 2.4%. The results of the study provide valuable information for operating and managing the Three Gorges Hydropower Station.

关 键 词：经验模态分解(EMD) Hilbert-Huang变换(HHT) 长江三峡水库 年平均流量预报 

分 类 号：P338.2[天文地球—水文科学]