东篱科研大数据发现系统（DRDS）

位置：成果数据库 > 期刊 > 期刊详情页

基于WA、ANN和水文频率分析法相结合的中长期水文预报模型的研究

ISSN号：1000-0852
期刊名称：水文
时间：0
页码：10-15
语言：中文
分类：P338.2[天文地球—水文科学;水利工程—水文学及水资源;天文地球—地球物理学]
作者机构：[1]南京大学地球科学与工程学院水科学系,江苏南京210093, [2]中水新华国际工程咨询公司,北京100053, [3]水利部黄河水利委员会水文局,河南郑州450001
相关基金：国家自然科学基金项目（40725010,40730635,40672160）;水利公益项目（2007SHZ1-24）
相关项目：水文地质学

关键词：中长期水文预报, 水文时间序列, 小波分析, 人工神经网络, 水文频率分析法, 不确定性, medium-and long-term hydrologic forecasting, hydrologic time series, wavelet analysis, artificial neural network, hydrologic frequency analysis, uncertainty

中文摘要：

针对传统中长期水文预报方法模拟预测结果精度低、未考虑水文不确定性因素的影响等问题，本文将小波分析（WA），人工神经网络（ANN）和水文频率分析法联合使用。建立了不确定性中长期水文预报模型：即在应用WA揭示水文序列变化特性的基础上。将原序列分为主序列和随机序列两部分。然后利用ANN对主序列进行模拟预测。对随机序列进行水文频率分析，最后将两部分结果叠加作为最终预测值。将该模型用于黄河河口地区作中长期水文预报，并与传统方法作对比，进行模型验证。结果显示：该模型能同时揭示序列的时、频结构和变化特性；预报值结果精度高；且合格率高；能定量分析和描述水文不确定性因素对预报结果的影响．可得到不同频率对应水文序列的模拟预测值。因此该模型的预报结果更加合理有效。对实际生产应用更具有指导意义。

英文摘要：

In order to improve the results of medium-and long-term hydrologic forecasting, and to analyze the influence of uncertain factors on the forecasting results, an stochastic medium-and long-term hydrologic forecasting model has been put forward based on WA, ANN, and hydrologic frequency analysis. The main idea of this model is as follows： first, analyze the multi-time scale characters of hydrologic time series by using WA method for understanding the internal structures of the series both in time and frequency domain, then get the main series in original series by DWT, and take the rest as random series. Then use ANN to simulate and forecast the main series, and the hydrologic frequency analysis method is used to get prediction of random series under different guarantee efficiency. Finally, stack the two as final forecasting results. The model was applied to the Estuary Area of the Yellow River for verifying. Results show that the model is of high precision, good percent of pass, can understand the characteristics of series meanwhile, and is able to describe the influence of uncertainty factors quantitatively. Therefore, the model is better than traditional forecasting models because of having more reasonable forecasting results and being more instructive for practical purpose.

同期刊论文项目