| 217 | 0 | 195 |
| 下载次数 | 被引频次 | 阅读次数 |
【目的】冲积河流阻力的准确计算在治河防洪工程中具有重要意义。传统阻力计算公式与现有机器学习方法仍存在诸多不足。为提升阻力模型的性能与泛化能力,建立一套基于深度学习的阻力计算方法,【方法】选择弗劳德数、体积含沙量、宽深比、径深比、年径流量和年输沙量等水文特征作为模型输入,构建一种基于深度森林的水流阻力计算模型。利用黄河下游水文站的实测数据进行模型训练与测试,并从时空泛化能力、迁移学习表现等方面综合评估模型。【结果】模型在测试集上的纳什效率(NSE)为0.785,平均绝对误差(MAE)为0.002,均方根误差(RMSE)为0.003,平均绝对百分比误差(MAPE)为14.618%。加入时空平均特征泛化模型后,模型的NSE从0.681 4提升至0.742 7, MAE从0.002 3降至0.002 1, RMSE从0.003 2降至0.002 8, MAPE从14.978%降至13.689%。在将模型迁移至全新的水沙条件下时, NSE最大降幅达65.35%, MAE、 RMSE和MAPE的最大升幅分别为100%、150%和123.98%。【结论】深度森林模型相比于传统的阻力公式和机器学习模型,在冲积河流一般条件下的阻力计算方面,能展现出更为优越的预测精度。通过引入大尺度时空平均特征,模型在不同水文站与水文时期的计算精度有效提升,泛化能力明显增强。然而,在应对特殊水沙情势时,模型仍存在性能波动,有时计算精度还超不过具有物理背景的糙率公式,亟需结合物理机制以增强其迁移学习能力,且在解决复杂环境下的黄河动床阻力计算问题时,应该注重同良好的传统公式相互印证这一环节。
Abstract：[Objective]Accurate calculation of flow resistance in alluvial rivers is of great significance for river regulation and flood control engineering. Conventional resistance formulas and existing machine learning method still have multiple limitations.To improve the performance and generalization ability of resistance models, a flow resistance estimation method based on deep learning is proposed. [Methods]Hydrological features, including Froude number, volumetric sediment concentration, width-todepth ratio, diameter-to-depth ratio, annual runoff, and annual sediment load, were selected as model inputs, and a flow resistance calculation model based on deep forest was established. The model was trained and tested using measured data from hydrological stations in the lower Yellow River, and comprehensively evaluated in terms of spatiotemporal generalization ability and transfer learning performance. [Results]The model achieved a Nash-Sutcliffe efficiency(NSE) of 0. 785, a mean absolute error(MAE) of 0. 002, a root mean square error(RMSE) of 0. 003, and a mean absolute percentage error(MAPE) of 14. 618% on the test dataset. After the incorporation of spatiotemporal average features, the NSE of the model increased from 0. 681 4 to 0. 742 7, the MAE decreased from 0. 002 3 to 0. 002 1, the RMSE dropped from 0. 003 2 to 0. 002 8, and the MAPE reduced from 14. 978% to 13. 689%. When the model was transferred to completely new water-sediment conditions, the maximum decline in NSE reached 65. 35%, and the maximum increases in MAE, RMSE, and MAPE were 100%, 150%, and 123. 98%, respectively. [Conclusion] Compared with traditional resistance formulas and machine learning method, the deep forest model demonstrates superior accuracy in predicting flow resistance under general conditions in alluvial rivers. By introducing large-scale spatiotemporal average features, the model's calculation accuracy across different hydrological stations and hydrological periods is effectively improved, and its generalization ability is significantly enhanced. However, under special water-sediment conditions, the model still shows performance fluctuations. In certain cases, its calculation accuracy is even lower than that of physically based roughness formulas. Therefore, it is urgent to incorporate physical mechanisms to enhance its transfer learning capability. When addressing the calculation of movable bed resistance in the Yellow River under complex environments, emphasis should be placed on mutual verification with reliable traditional formulas.
[1] JIANG R J, LU H, YANG K, et al. Substantial increase in future fluvial flood risk projected in China’ s major urban agglomerations[J]. Communications Earth&Environment, 2023, 4(1):389.
[2] MOTAMEDI A, AFZALIMEHR H, SINGH V P. Estimation of friction factor in open channels[J]. Journal of Hydrologic Engineering, 2010, 15(3):249-254.
[3] BELLOS V, NALBANTIS I, TSAKIRIS G. Friction modeling of flood flow simulations[J]. Journal of Hydraulic Engineering, 2018, 144(12):04018073.
[4]翟明杰,焦彩虹,庞慧霞,等.糙率及地面沉降对河道行洪能力影响的研究[J].水利发展研究, 2024, 24(11):29-34.ZHAI Mingjie, JIAO Caihong, PANG Huixia, et al. Discussion on the influence of roughness and land subsidence on flood discharge capacity of watercourse[J]. Water Resources Development Research,2024, 24(11):29-34.
[5] KUMAR S, KHUNTIA J R, KHATUA K K. Prediction of flow resistance in an open channel over movable beds using artificial neural network[J]. Journal of Hydrologic Engineering, 2021, 26(5):04021015.
[6] STAUDT F, MULLARNEY J C, PILDITCH C A, et al. Effects of grain-size distribution and shape on sediment bed stability, near-bed flow and bed microstructure[J]. Earth Surface Processes and Landforms, 2019, 44(5):1100-1116.
[7]穆祥鹏,张哲琦,李军强,等.纳米胶在混凝土输水工程中的减阻降糙特性[J].南水北调与水利科技(中英文), 2025, 23(2):473-480.MU X P, ZHANG Z Q, LI J Q, et al. Preliminary study on the drag reduction and roughness reduction properties of Nano-Rubber applications in concrete water conveyance engineering[J]. South-toNorth Water Transfers and Water Science&Technology, 2025, 23(2):473-480.
[8] LIU J Z, ZHANG J L, HUANG Z, et al. Prediction of movable bed resistance in open channels with dune forms incorporating the effects of flow separation and water temperature[J]. Journal of Hydrology,2024, 634:131073.
[9] RAUDKIVI A J. Analysis of resistance in fluvial channel[J]. Journal of the Hydraulics Division, 1967, 93(5):73-84.
[10] YALIN S. On the average velocity of flow over a movable bed[J]. La Houille Blanche, 1964(1):45-51.
[11] VAN RIJN L C. Sediment transport, part III:bed forms and alluvial roughness[J]. Journal of Hydraulic Engineering, 1984, 110(12):1733-1754.
[12] WHITE W R, PARIS E, BETTESS R, et al. The frictional characteristics of alluvial streams:A new approach[J]. Proceedings of the Institution of Civil Engineers, 1980, 69(3):737-750.
[13] ALAM A M Z, KENNEDY J F. Friction factors for flow in sand-bed channels[J]. Journal of the Hydraulics Division, 1969, 95(6):1973-1992.
[14] YEN B C. Open channel flow resistance[J]. Journal of Hydraulic Engineering, 2002, 128(1):20-39.
[15] GUO S L, WEN Y H, ZHANG X Q, et al. Research on precipitation prediction based on a complete ensemble empirical mode decomposition with adaptive noise-long short-term memory coupled model[J]. Water Supply, 2022, 22(12):9061-9072.
[16] FENG Z K, NIU W J, WAN X Y, et al. Hydrological time series forecasting via signal decomposition and twin support vector machine using cooperation search algorithm for parameter identification[J].Journal of Hydrology, 2022, 612:128213.
[17] ZHANG X Q, YANG Y. Suspended sediment concentration forecast based on CEEMDAN-GRU model[J]. Water Supply, 2020, 20(5):1787-1798.
[18] ZHI W, APPLING A P, GOLDEN H E, et al. Deep learning for water quality[J]. Nature Water, 2024, 2(3):228-241.
[19] NIAZKAR M, TALEBBEYDOKHTI N, AFZALI S H. Novel grain and form roughness estimator scheme incorporating artificial intelligence models[J]. Water Resources Management, 2019, 33:757-773.
[20] MILUKOW H A, BINNS A D, ADAMOWSKI J, et al. Estimation of the Darcy-Weisbach friction factor for ungauged streams using Gene Expression Programming and Extreme Learning Machines[J]. Journal of Hydrology, 2019, 568:311-321.
[21] KARBASI M, GHASEMIAN M, JAMEI M, et al. Developing extended and unscented Kalman Filter-based neural networks to predict cluster-induced roughness in gravel bed rivers[J]. Water Resources Management, 2024, 38(8):3023-3048.
[22] SAGHEBIAN S M, ROUSHANGAR K, OZGUR KIRCA V S, et al.Modeling total resistance and form resistance of movable bed channels via experimental data and a kernel-based approach[J]. Journal of Hydroinformatics, 2020, 22(3):528-540.
[23] XIA X H, DONG J W, WANG M H, et al. Effect of water-sediment regulation of the Xiaolangdi reservoir on the concentrations,characteristics, and fluxes of suspended sediment and organic carbon in the Yellow River[J]. Science of the Total Environment, 2016,571:487-497.
[24] YANG R Y, PENG Y, ZHANG H W, et al. Estimation of flow resistance in natural rivers based on deep forest[J]. Earth Science Informatics, 2025, 18(1):1-16.
[25]侯琳,张红武,李琳琪,等.黄河水流阻力及河床比降计算方法研究[J].水利学报, 2024, 55(9):1020-1032.HOU L, ZHANG H W, LI L Q, et al. Study on calculation of flow resistance and riverbed slope of the Yellow River[J]. Journal of Hydraulic Engineering, 2024, 55(9):1020-1032.
[26]张红武,张罗号,彭昊,等.冲积河流糙率由来与计算方法研究[J].水利学报, 2020, 51(7):774-787.ZHANG H W, ZHANG L H, PENG H, et al. Research on cognition and calculation method of alluvial river roughness[J]. Journal of Hydraulic Engineering, 2020, 51(7):774-787.
[27]李琳琪,张红武,侯琳,等.河流动床阻力影响参数分析及其计算研究[J].水利学报, 2023, 54(12):1496-1506.LI L Q, ZHANG H W, HOU L, et al. Parameter analysis and calculation method of moving bed resistance[J]. Journal of Hydraulic Engineering, 2023, 54(12):1496-1506.
[28]张罗号.黄河河槽糙率异常原因及其解决途径[J].水利学报,2012, 43(11):1261-1270.ZHANG L H. Reasons for the abnormal channel roughness of the Yellow River and the solution to its problems[J]. Journal of Hydraulic Engineering, 2012, 43(11):1261-1270.
[29]张罗号,卜海磊,田依林.细沙河流床沙中径计算公式研究[J].人民黄河, 2011, 33(12):38-41.ZHANG L H, BU H L, TIAN Y L. Study on the formula of medium size of fine sand of bed material[J]. Yellow River, 2011, 33(12):38-41.
[30] CHANG Y L. Laboratory investigation of flume traction and transportation[J]. Transactions of the American Society of Civil Engineers, 1939, 104(1):1246-1284.
[31]张红武,江恩惠,白咏梅,等.黄河高含沙洪水模型的相似律[M].郑州:河南科学技术出版社, 1994.ZHANG H W, JIANG E H, BAI Y M, et al. The Scale Law of Model Similarity of Hyperconcentrated Flow[M]. Zhengzhou:Henan Science and Technology Press, 1994.
[32]彭昊,黄河清,张红武,等.体现水流与沙粒摩擦阻力相对作用的糙率公式[J].泥沙研究, 2022, 47(2):1-8.PENG H, HUANG H Q, ZHANG H W, et al. A roughness formula based on the relativity of flow intensity and grain roughness[J].Journal of Sediment Research, 2022, 47(2):1-8.
[33] HOU L, ZHANG H W, LI L Q, et al. A method for calculating water demand for sediment transport based on the principles of river dynamics[J]. Water, 2023, 15(19):3514.
[34] LIMERINOS J T. Determination of the Manning Coefficient from Measured Bed Roughness in Natural Channels[M]. Washington, D.C.:US Government Printing Office, 1970.
[35] ZHOU Z H, FENG J. Deep forest[J]. National Science Review,2019, 6(1):74-86.
[36] DAVID G C L, WOHL E, YOCHUM S E, et al. Controls on spatial variations in flow resistance along steep mountain streams[J]. Water Resources Research, 2010, 46(3):1-21.
[37] WANG Z Y, QI P, MELCHING C S. Fluvial hydraulics of hyperconcentrated floods in Chinese rivers[J]. Earth Surface Processes and Landforms, 2009, 34(7):981-993.
[38] DAVIES T R, SUTHERLAND A J. Resistance to flow past deformable boundaries[J]. Earth Surface Processes, 1980, 5(2):175-179.
[39] GARG A, TAI K. Comparison of statistical and machine learning methods in modelling of data with multicollinearity[J]. International Journal of Modelling, Identification and Control, 2013, 18(4):295-312.
[40] CHAN J Y L, LEOW S M H, BEA K T, et al. Mitigating the multicollinearity problem and its machine learning approach:a review[J]. Mathematics, 2022, 10(8):1283.
[41] FARRAR D E, GLAUBER R R. Multicollinearity in regression analysis:The problem revisited[J]. The Review of Economic and Statistics, 1967, 49(1):92-107.
[42]侯琳,张红武.黄河下游河道摩阻计算方法研究[J].水利学报, 2025, 56(1):63-72.HOU L, ZHANG H W. Research on the calculation method of frictional resistance in the lower reaches of the Yellow River[J].Journal of Hydraulic Engineering, 2025, 56(1):63-72.
[43] YARAHMADI M B, PARSAIE A, SHAFAI-BEJESTAN M, et al.Estimation of Manning roughness coefficient in alluvial rivers with bed forms using soft computing models[J]. Water Resources Management, 2023, 37(9):3563-3584.
[44] MIR A A, PATEL M. Machine learning approaches for adequate prediction of flow resistance in alluvial channels with bedforms[J].Water Science&Technology, 2024, 89(2):290-318.
[45] HRACHOWITZ M, SAVENIJE H H G, BLÖSCHL G, et al. A decade of predictions in ungauged basins(PUB):A review[J].Hydrological Sciences Journal, 2013, 58(6):1198-1255.
[46]张红武,赵晨苏,张罗号,等.内蒙古自治区水利行业数字孪生总体构架方案[R].北京:清华大学, 2024.ZHANG H W, ZHAO C S, ZHANG L H, et al. Overall Framework for Digital Twins in the Water Resources Sector of the Inner Mongolia Autonomous Region[R]. Beijing:Tsinghua University, 2024.
[47]麻妍妍,夏军强,张晓雷.现有动床阻力计算公式验证与比较[J].武汉大学学报(工学版), 2017, 50(4):481-486.MA Y Y, XIA J Q, ZHANG X L. Verification and comparison of flow resistance formulae over movable beds[J]. Engineering Journal of Wuhan University, 2017, 50(4):481-486.
[48]杨润祎,张红武.基于涡团模式的紊流流速与阻力系数的统一规律[J].中国科学:物理学力学天文学, 2026, 56(3):234711.YANG R Y, ZHANG H W. Unified law of turbulent flow velocity and drag coefficient based on turbulent eddy model[J]. Scientia Sinica(Physica, Mechanica&Astronomica), 2026, 56(3):234711.
[49]杨润祎,张红武,朱金峰.河流泥沙颗粒分析规程修编对悬移质级配分布测算的影响研究[J].水利学报, 2025, 56(3):317-327.YANG R Y, ZHANG H W, ZHU J F. Effect of the revision of technical standard for determination of sediment particle size on the estimation of particle size distribution[J]. Journal of Hydraulic Engineering, 2025, 56(3):317-327.
[50]张红武,李琳琪,彭昊,等.基于流域高质量发展目标的黄河相关问题研究[J].水利水电技术(中英文), 2021, 52(12):60-68.ZHANG Hongwu, LI Linqi, PENG Hao, et al. Study on related problems of the Yellow River based on the goal of high-quality development of the basin[J]. Water Resources and Hydropower Engineering, 2021, 52(12):60-68.
[51]张红武,张清.黄河水流挟沙力的计算公式[J].人民黄河,1992, 14(11):7-9.ZHANG H W, ZHANG Q. Formula of sediment carrying capocity of the Yellow River[J]. Yellow River, 1992, 14(11):7-9.
[52]张罗号,卜海磊.黄河下游糙率奇小的原因分析[J].人民黄河,2010, 32(12):84-86.ZHANG L H, BU H L. Analysis of the extremely low roughness coefficient in the lower reaches of the Yellow River[J]. Yellow River, 2010, 32(12):84-86.
基本信息:
DOI：10.13928/j.cnki.wrahe.2026.01.019
中图分类号:TP18;TV143
引用信息:
[1]杨润祎,张红武.基于深度学习的黄河下游水流阻力计算研究[J].水利水电技术(中英文),2026,57(01):249-262.DOI:10.13928/j.cnki.wrahe.2026.01.019.
基金信息:
国家重点研发计划项目(2023YFC3208603)
2025-08-25
2025-08-25
2025-08-25