Prediction and Analysis of Long-Term Deformation for Gravity Dams Integrating Dam Temperature Field Information and Adaptive Feature Weights

ZHENG Yuhe; TANG Xiaosong; KE Qin; ZANG Hanghang

doi:10.14188/j.1671-8844.2026.0056

您当前的位置：

首页 >

文章列表页 >

Prediction and Analysis of Long-Term Deformation for Gravity Dams Integrating Dam Temperature Field Information and Adaptive Feature Weights

更新时间：2026-10-15

- Prediction and Analysis of Long-Term Deformation for Gravity Dams Integrating Dam Temperature Field Information and Adaptive Feature Weights
- Engineering Journal of Wuhan University Pages: 1-14(2026)
- 作者机构：
  
  1.武汉大学水资源工程与调度全国重点实验室，湖北武汉 430072
  2.武汉大学水利水电学院水工程风险与防灾研究所，湖北武汉 430072
- 作者简介：
- 基金信息：
- DOI：10.14188/j.1671-8844.2026.0056
  CLC： TU 457
- Online First：15 October 2026，
- 稿件说明：
移动端阅览
ZHENG Yuhe, TANG Xiaosong, KE Qin, et al. Prediction and Analysis of Long-Term Deformation for Gravity Dams Integrating Dam Temperature Field Information and Adaptive Feature Weights[J/OL]. Engineering Journal of Wuhan University, 2026, 1-14.
DOI：

ZHENG Yuhe, TANG Xiaosong, KE Qin, et al. Prediction and Analysis of Long-Term Deformation for Gravity Dams Integrating Dam Temperature Field Information and Adaptive Feature Weights[J/OL]. Engineering Journal of Wuhan University, 2026, 1-14. DOI： 10.14188/j.1671-8844.2026.0056.

摘要

针对以往重力坝长期变形预测研究中温度效应表征不完善、模型特征与模型参数独立优化等问题，本文提出一种融合坝体温度场信息与自适应特征权重的智能预测方法。首先，通过引入坝内温度计实测数据完善温度特征集。然后，采用卷积神经网络-双向长短期记忆网络-注意力机制（Convolutional Neural Network-Bidirectional Long Short-Term Memory-Attention

CNN-BiLSTM-Attention）构建大坝变形混合预测模型，逐层提取多因子局部特征、建模双向时序依赖并聚焦关键时段。接着，采用北方苍鹰优化（Northern Goshawk Optimization

NGO）算法协同优化特征权重与模型超参数，实现特征评估与模型结构协同寻优。最后，基于核密度估计方法量化模型预测的不确定性。实例验证表明，得益于坝体温度场信息的引入与自适应特征-模型参数的协同优化机制，本文方法实现模型精度提升和得到更丰富的特征信息。

Abstract

Previous research on long-term deformation prediction of gravity dams has limitations

including inadequate characterization of temperature effects and separate optimization of model features and hyperparameters. To address these issues

this study proposes an intelligent prediction method that integrates dam temperature field information and adaptive feature weighting. First

measured data from embedded thermometers is incorporated to enrich the temperature feature set. Second

a hybrid prediction model for dam deformation is developed based on a Convolutional Neural Network–Bidirectional Long Short-Term Memory–Attention (CNN-BiLSTM-Attention) framework. This model is designed to progressively extract local features from multiple influencing factors

capture bidirectional temporal dependencies

and identify critical time steps.Subsequently

the Northern Goshawk Optimization (NGO) algorithm is employed to simultaneously optimize both feature weights and model hyperparameters

thereby achieving a synergy between feature evaluation and model structure.Finally

prediction uncertainty is quantified using a kernel density estimation approach. A case study verifies that by integrating dam temperature field information and enabling collaborative optimization between feature weights and model hyperparameters

the proposed method effectively enhances prediction accuracy while extracting more informative features.

关键词

Keywords

references

顾冲时 , 苏怀智 , 刘何稚 . 大坝服役风险分析与管理研究述评 [J ] . 水利学报 , 2018 , 49 ( 01 ): 26 - 35 .

GUChongshi , SUHuaizhi , LIUHezhi . Reviewonserviceriskanalysisofdamengineering [J ] . JournalofHydraulicEngineering , 2018 , 49 ( 01 ): 26 - 35 .

YaoK , WenZ , ShaoC , etal . Amultisourcedata‐drivenmonitoringmodelforassessingconcretedambehavior [J ] . Computer-AidedCivil&InfrastructureEngineering , 2024 , 39 ( 23 ): 3595 - 3609 .

刘健 , 王继敏 , 杨强 , 等 . 混凝土大坝变形监测异常值识别方法及应用 [J ] . 水利水电科技进展 , 2025 , 45 ( 06 ): 77 - 84 .

LIUJian , WANGJimin , YANGQiang , etal . Methodforidentificationofoutliersinconcretedamdeformationmonitoringanditsapplication [J ] . AdvancesinScienceandTechnologyofWaterResources , 2025 , 45 ( 06 ): 77 - 84 .

彭绍才 , 郑栋 , 韦国书 . 施工期乌东德水电站左岸尾水边坡实测变形及支护受力特性 [J ] . 武汉大学学报(工学版) , 2022 , 55 ( 02 ): 131 - 143 .

PENGShaocai , ZHENGDong , WEIGuoshu . Deformationandsupportingforcecharacteristicsofleftbanktail‐waterslopeofWudongdeHydropowerStationduringconstructionperiod [J ] . EngineeringJournalofWuhanUniversity , 2022 , 55 ( 02 ): 131 - 143 .

唐贤琪 , 施玉群 , 杨海云 , 等 . 基于贝叶斯网络推理的大坝风险评估 [J ] . 武汉大学学报(工学版) , 2024 , 57 ( 02 ): 152 - 158 .

TANGXianqi , SHIYuqun , YANGHaiyun , etal . DamriskassessmentbasedonBayesiannetworkinference [J ] . EngineeringJournalofWuhanUniversity , 2024 , 57 ( 02 ): 152 - 158 .

吴中如 . 水工建筑物安全监控理论及其应用 [Z ] . 河海大学.北京 : 高等教育出版社 , 2003:45-58

TatinM , BriffautM , DufourF , etal . Statisticalmodellingofthermaldisplacementsforconcretedams:Influenceofwatertemperatureprofileanddamthicknessprofile [J ] . EngineeringStructures , 2018 ,165(jun. 15 ): 63 - 75 .

WangS , XuY , GuC , etal . Hystereticeffectconsideredmonitoringmodelforinterpretingabnormaldeformationbehaviorofarchdams:Acasestudy [J ] . StructuralControlandHealthMonitoring , 2019 , 26 ( 10 ): 1 - 20 .

TatinM , BriffautM , DufourF , SimonA , FabreJP . Thermaldisplacementsofconcretedams:accountingforwatertemperatureinstatisticalmodels . EngineeringStructures , 2015 ( 91 ): 26 – 39 .

PrakashG , SadhuA , NarasimhanS , etal . Initialservicelifedatatowardsstructuralhealthmonitoringofaconcretearchdam [J ] . StructuralControlandHealthMonitoring , 2018 , 25 ( 1 ): 1 - 19 .

徐伟 , 何金平 . 基于多尺度小波分析的大坝变形自回归预测模型 [J ] . 武汉大学学报(工学版) , 2012 , 45 ( 03 ): 285 - 289 .

XUWei , HEJinping . Forecastmodelofdamdeformationbasedonmulti-scalewaveletanalysisandautoregressivemethod [J ] . EngineeringJournalofWuhanUniversity , 2012 , 45 ( 03 ): 285 - 289 .

胡雨菡 , 包腾飞 , 朱征 , 等 . 基于IABC-FCM-RVM算法的拱坝变形预测模型 [J ] . 武汉大学学报(工学版) , 2020 , 53 ( 12 ): 1055 - 1064 .

HUYuhan , BAOTengfei , ZHUZheng , etal . IABC-FCM-RVMbasedpredictionmodelofarchdamdeformation [J ] . EngineeringJournalofWuhanUniversity , 2020 , 53 ( 12 ): 1055 - 1064 .

冯宇扬 , 李登华 , 方博雅 , 等 . 关于大坝监测数据质量评价因子及算法研究 [J ] . 现代电子技术 , 2025 , 48 ( 02 ): 90 - 96 .

FENGYuyang , LIDenghua , FANGBoya , etal . Researchonevaluationfactorsandalgorithmsofdammonitoringdataquality [J ] . ModernElectronicsTechnique , 2025 , 48 ( 02 ): 90 - 96 .(权重分离)

赵明华 , 刘建华 , 陈炳初 , 等 . 边坡变形及失稳的变权重组合预测模型 [J ] . 岩土力学 , 2007 , 28 ( S1 ): 553 - 557 .

ZHAOMinghua , LIUJianhua , CHENBingchu , etal . Variable-weightcombinationforecastingmodelofslopedeformationandinstability [J ] . RockandSoilMechanics , 2007 , 28 ( S1 ): 553 - 557 .

赵文光 , 邱焕峰 . BSG高碾压混凝土拱坝温度场仿真分析 [J ] . 武汉大学学报(工学版) , 2020 , 53 ( S1 ): 299 - 303 .

ZHAOWenguang , QIUHuanfeng . SimulationanalysisoftemperaturefieldofBSGhighRCCarchdams [J ] . EngineeringJournalofWuhanUniversity , 2020 , 53 ( S1 ): 299 - 303 .

常泽 , 王立成 . 混凝土内部温度场的细观数值模拟研究 [J ] . 武汉大学学报(工学版) , 2017 , 50 ( 06 ): 938 - 942 .

CHANGZe , WANGLicheng . Mesoscopicnumericalsimulationofthetemperaturefieldofconcrete [J ] . EngineeringJournalofWuhanUniversity , 2017 , 50 ( 06 ): 938 - 942 .

周凯 , 焦龄霄 , 胡志坚 , 等 . 基于注意力机制的CNN-BiLSTM操作票自动校核方法 [J ] . 武汉大学学报(工学版) , 2023 , 56 ( 09 ): 1114 - 1123 .

ZHOUKai , JIAOLingxiao , HUZhijian , etal . AutomaticverificationmethodofoperationticketbasedonattentionmechanismofCNN-BiLSTM [J ] . EngineeringJournalofWuhanUniversity , 2023 , 56 ( 09 ): 1114 - 1123 .

韩冰青 , 熊培淞 . 基于改进NGO算法的LEACH多跳路由优化方法 [J ] . 计算机工程与设计 , 2025 , 46 ( 11 ): 3167 - 3173 .

HANBingqing , XIONGPeisong . OptimizedmethodforLEACHmulti-hoproutingbasedonimprovedNGOalgorithm [J ] . ComputerEngineeringandDesign , 2025 , 46 ( 11 ): 3167 - 3173 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Study on tracking inversion of rockfill dams with prior information supervision

Seismic response analysis of concrete gravity dam based on acoustic‑structural coupling method

Analysis of the working state and stability safety of No. 2 mud dam in Xiaomaliu Tailing Pond

Analysis of the influence of non‑homogeneous overburden foundation on the stress and deformation of concrete gravity dam

The deep anti-sliding stability analysis of gravity dam based on Hoek-Brown failure criterion

Related Author

WANG Jiawei

MA Gang

LI Xiang

AI Zhitao

XU Jianhua

ZHOU Wei

LIU Aobo

WANG Gaohui

Related Institution

Hubei Energy Group Loushui Hydropower Co.， Ltd.

Institute of Water Engineering Sciences， Wuhan University

Key Laboratory of Rock Mechanics in Hydraulic Structural Engineering of Ministry of Education， Wuhan University

State Key Laboratory of Water Resources and Hydropower Engineering Science， Wuhan University

Zhongye Changtian International Engineering Co.Ltd.

AI问答

⁰