摘要
内河航道等级对流场时空域水文水动力要素响应复杂,寻找基于响应机理的等级识别模型是水库精细化管理前提。借助赣江峡江—新干段分布式水动力模型,开展逐日水文水动力模拟;耦合概率分布和机器学习算法,构建航道等级智能识别模型,评价航道等级时空分布。结果表明,水文水动力模拟精度高,模拟和实测日水位决定系数(R^(2))≥0.84、Nash系数≥0.75。Stable分布是河宽、水深、弯曲半径等预报变量最优分布,特征指数、对称、尺度和位置参数分别在1.21~1.61、-1.00~1.00、0.31~108.32、3.51~969.11变化。人工神经网络等是等级识别适宜算法,建模和验证期R^(2)、Nash系数≥0.98。航道等级基本保持三级以上,三级航道保证率最小值出现在峡江坝后和枯水期,是航段管理重点区间和时段。
The classification of inland waterways is influenced by complex spatiotemporal responses of hydrological and hydrodynamic factors within the flow field.Developing a classification model based on such response mechanisms is a prerequisite for refined reservoir management.In this study,a distributed hydrodynamic model was employed to the Xiajiang-Xingan section of the Ganjiang River to simulate daily hydrological and hydrodynamic processes.Inland waterway classification model was built by coupling the probability distribution of hydrological and hydrodynamic elements and machine learning alg orithms.The inland waterway classes'spatiotemporal distribution was assessed using the model.It indicated that the hydro-hydrodynamic simulation presented a relatively high accuracy,proved by the determination coefficient(R^(2))≥0.84 and Nash-Sutcliffe efficiency coefficient≥0.75.Stable distribution was detected as the best fitted one for the river width,water depth,and radius of curvature,the characteristic exponent,symmetry,scale,and location parameters of which ranged at 1.21~1.61,-1.00~1.00,0.31~108.32 and 3.51~969.11.Artificial neural networks and related alg orithms were shown to be suitable for waterway grade recognition,with both modeling and validation periods achieving R^(2)and Nash coefficients≥0.98.Waterway grades were generally maintained at Grade Ⅲ or higher,with the lowest assurance rate of Grade Ⅲ occurring downstream of the Xiajiang Dam during the dry season,which emphasized key regions and periods for targeted channel management.
作者
章雨铖
白桦
曹裕霖
肖文昌
戈晓斌
杨筱筱
温珍玉
李斌
ZHANG Yucheng;BAI Hua;CAO Yulin;XIAO Wenchang;GE Xiaobin;YANG Xiaoxiao;WEN Zhenyu;LI Bin(Jiangxi Key Laboratory of Water Resources Allocation and Efficient Utilization of Nanchang Institute of Technology,330099,Nanchang,PRC;Jiangxi Provincial Port&Waterway Construction Investment Group Co.,Ltd.,330200,Nanchang,PRC;Jiangxi Provincial Hydrological Monitoring Center,330038,Nanchang,PRC)
出处
《江西科学》
2025年第4期651-659,共9页
Jiangxi Science
基金
江西省重点研发计划项目(20212BBG71014)
江西省水利厅科技项目(202425YBKT15)
江西省港航建设投资集团有限公司科技项目(2023-YJY-RD02)。
关键词
水文学
航道等级
机器学习
水动力模拟
边际分布
hydrology
ranking of waterway
machine learning
hydrodynamic simulation
marginal distribution
作者简介
章雨铖(2001-),男,硕士研究生,主要从事水文学及水资源研究;通信作者:白桦(1986-),男,博士,副教授,主要从事水文学及水资源研究。E-mail:baihua1985@126.com。
