|
| 引汉济渭工程水源区月径流预报模型研究 |
| Study on Monthly Runoff Forecasting Models for the Water Source Area of the Hanjiang-to-Weihe River Diversion Project |
| 投稿时间：2024-12-05 修订日期：2024-12-31 |
| DOI： |
| 中文关键词: 月径流预报模型 引汉济渭工程 分解-重构 变分模态分解 |
| 英文关键词:Monthly Runoff Forecasting Model Hanjiang-to-Weihe River Diversion Project Decomposition-Reconstruction Variational Mode Decomposition (VMD) |
| 基金项目:2023年陕西省博士后科研项目资助(2023BSHGZZHQYXMZZ51) |
|
| 摘要点击次数: 116 |
| 全文下载次数: 0 |
| 中文摘要: |
| 径流预测的准确性对于水资源管理至关重要。针对径流的非线性、突变和非平稳性特征,本文提出了一种基于分解-重构技术的混合模型,以提高引汉济渭工程水源区——三河口水库月入库径流的预测精度。通过变分模态分解(VMD)和集总经验模态分解(EEMD)对原始径流序列进行分解,获得多个固有模态函数(IMFs),并采用梯度提升回归树(GBRT)对各IMF进行预测。研究对比了支持向量回归(SVR)、GBRT、EEMD-GBRT和VMD-GBRT模型的预测性能,并分析了它们在Nash-Sutcliffe效率系数(NSE)、均方根误差(RMSE)、平均绝对误差(MAE)及平均绝对百分比误差(MAPE)四个指标上的表现。结果表明,VMD-GBRT模型在捕捉流量变化趋势和预测精度方面表现最佳。综合分析证明,VMD-GBRT模型是一种有效的径流预测方法,为水资源管理提供了可靠的技术支持。 |
| 英文摘要: |
| Accurate runoff forecasting is crucial for effective water resource management. To address the nonlinear, abrupt, and non-stationary characteristics of runoff, this study proposes a hybrid model based on decomposition-reconstruction techniques to enhance the prediction accuracy of monthly inflow runoff into the Sanhekou Reservoir, a water source area for the Hanjiang-to-Weihe River Diversion Project. The original runoff series was decomposed into multiple intrinsic mode functions (IMFs) using Variational Mode Decomposition (VMD) and Ensemble Empirical Mode Decomposition (EEMD). Gradient Boosting Regression Tree (GBRT) was employed to predict each IMF. The study compared the predictive performance of Support Vector Regression (SVR), GBRT, EEMD-GBRT, and VMD-GBRT models, analyzing their results based on metrics such as the Nash-Sutcliffe Efficiency Coefficient (NSE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE). The results demonstrate that the VMD-GBRT model outperformed others in capturing runoff variation trends and achieving higher prediction accuracy. Comprehensive analysis confirms that the VMD-GBRT model is an effective runoff forecasting method, providing reliable technical support for water resource management. |
| 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
