摘要
提高径流预测的适用性对水资源的合理开发和高效利用具有重要意义。针对传统点预测方法无法有效描述预测结果的不确定性问题,提出了基于VMD-GRU和非参数核密度估计的月径流区间预测方法。首先采用变分模态分解(VMD)将月径流序列分解为一系列相对平稳的子序列,然后利用门控循环单元(GRU)分别预测各子序列,叠加得到最终的点预测结果,最后在点预测的基础上,应用非参数核密度估计进行月径流区间预测,并将提出的VMD-GRU模型与GRU、极点对称模态分解-门控循环单元(ESMD-GRU)和完全集合经验模态分解-门控循环单元(CEEMDAN-GRU)模型进行对比。结果表明,该模型点预测精度明显高于其他模型,同时非参数核密度估计为径流区间预测提供了合理的波动范围,可为管理决策提供参考。
Improving the applicability of runoff prediction is of great significance to the rational development and efficient utilization of water resources. Aiming at the problem that the traditional point prediction methods cannot effectively describe the uncertainty of prediction results, a monthly runoff interval prediction method based on VMD-GRU and nonparametric kernel density estimation was proposed. Firstly, the monthly runoff sequence was decomposed into a series of relatively stable sub-sequences using variational mode decomposition(VMD). And then each sub-sequence was separately predicted by the gated recurrent unit(GRU), and the final point prediction result was obtained by superposition. Finally, on the basis of point prediction, nonparametric kernel density estimation was used to predict monthly runoff interval. The proposed VMD-GRU model was compared with GRU, ESMD-GRU and CEEMDAN-GRU models. The results show that the point prediction accuracy of this model is significantly higher than other models. At the same time, nonparametric kernel density estimation provides a reasonable fluctuation range for runoff interval prediction, which can provide a reference for management decision-making.
作者
徐冬梅
王亚琴
王文川
XU Dong-mei;WANG Ya-qin;WANG Wen-chuan(College of Water Resources,North China University of Water Resources and Electric Power,Zhengzhou 450046,China)
出处
《水电能源科学》
北大核心
2022年第6期1-5,共5页
Water Resources and Power
基金
河南省科技攻关项目(202102310259,202102310588)
国家自然科学基金项目(51509088)
河南省高校科技创新团队项目(18IRTSTHN009)。
作者简介
徐冬梅(1977-),女,博士、教授、博导,研究方向为水资源系统分析,E-mail:xudongmei@ncwu.edu.cn;通讯作者:王文川(1976-),男,博士、教授、博导,研究方向为工程水文与水资源利用,E-mail:wangwen1621@163.com。
