| 基于BP-DEMATEL算法的冰情预报因子敏感性分析 |
| |
| 引用本文： | 孙亚翡,王涛,路锦枝,周中元.基于BP-DEMATEL算法的冰情预报因子敏感性分析[J].水利学报,2022,53(9):1083-1091. |
| |
| 作者姓名： | 孙亚翡 王涛 路锦枝 周中元 |
| |
| 作者单位： | 中国水利水电科学研究院 流域水循环与调控国家重点实验室, 北京 100038;中国水利水电科学研究院 流域水循环与调控国家重点实验室, 北京 100038;清华大学 水利水电工程系, 北京 100084 |
| |
| 基金项目： | 国家自然科学基金项目(U2243221,51979291,52009144);中国水科院科研专项(HY0145B032021,HY110145B0012021);流域水循环模拟与调控国家重点实验室自由探索课题(SKL2022TS04) |
| |
| 摘 要： | 冰情发展受到水文、气象、水力、河道条件和人类活动等多因子相互作用的影响,获得各影响因子的权重关系是进一步明晰冰情发展演变规律和提高预报精度的基础。本研究提出了基于BP-DEMATEL算法的冰情预报因子敏感性分析模型,应用于黄河内蒙古河段巴彦高勒水文站流凌、封河及开河冰情影响因子的分析中,得到影响冰情演变各因子的权重和不同因子之间的相互耦合关系,明确了冰情预报中关键的影响因子。采用不同权重的因子开展流凌、封河和开河的预报,结果显示利用权重值大、相关性强的预报因子开展冰情预报的预报值与实测值吻合较好。因此本研究提出的BP-DEMATEL模型开展冰情预报因子敏感性分析能够得到合理的权重值。
|
| 关 键 词： | 冰情 相关因子 BP-DEMATEL 神经网络 冰情预报 |
| 收稿时间： | 2022/3/31 0:00:00 |
Sensitivity analysis of BP-DEMATEL model to control parameters of ice processes |
| |
| SUN Yafei,WANG Tao,LU Jinzhi,ZHOU Zhongyuan.Sensitivity analysis of BP-DEMATEL model to control parameters of ice processes[J].Journal of Hydraulic Engineering,2022,53(9):1083-1091. |
| |
| Authors: | SUN Yafei WANG Tao LU Jinzhi ZHOU Zhongyuan |
| |
| Affiliation: | State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, China;State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, China;Department of Hydraulic Engineering, Tsinghua University, Beijing 100084, China |
| |
| Abstract: | The development of ice condition is affected by the interaction of various factors, like hydrology, weather and boundary conditions. The weight of influencing parameters is the basis improving the forecast accuracy and understanding the evolution mechanism of river ice. In this study, the BP -DEMATEL model is presented to analyzing the inner links of various key parameters affecting the date of ice run, freeze-up and break-up at Bayangaole station in the Yellow River. These parameters with different importance are used to forecast the river ice conditions. The result shows that the predicted ice condition using parameters with strong correlation provides better agreement with the observation data. Therefore, the BP-DEMATEL model proposed in this study can improve the ice condition forecasting. |
| |
| Keywords: | ice condition related factors BP-DEMATEL neural network ice forecasting |
|
| 点击此处可从《水利学报》浏览原始摘要信息 |
|
点击此处可从《水利学报》下载全文 |
|
