Deformation regression analysis of arch dam based on mixed random forest and shap methods

LIANG Haijing; CHENG Jing; YANG Junyi; YU Lei; DAI Miaolin

doi:10.14188/j.1671-8844.2025.0354

您当前的位置：

首页 >

文章列表页 >

Deformation regression analysis of arch dam based on mixed random forest and shap methods

更新时间：2026-10-15

- Deformation regression analysis of arch dam based on mixed random forest and shap methods
- Engineering Journal of Wuhan University Pages: 1-10(2026)
- 作者机构：
  
  1.河海大学水利水电学院，江苏南京 210098
  2.水利部黄河流域水治理与水安全重点实验室，河南郑州 450003
  3.贵州省大坝安全监测中心，贵州贵阳 550002
- 作者简介：
- 基金信息：
- DOI：10.14188/j.1671-8844.2025.0354
  CLC： TV698.1
- Online First：15 October 2026，
- 稿件说明：
移动端阅览
LIANG Haijing, CHENG Jing, YANG Junyi, et al. Deformation regression analysis of arch dam based on mixed random forest and shap methods[J/OL]. Engineering Journal of Wuhan University, 2026, 1-10.
DOI：

LIANG Haijing, CHENG Jing, YANG Junyi, et al. Deformation regression analysis of arch dam based on mixed random forest and shap methods[J/OL]. Engineering Journal of Wuhan University, 2026, 1-10. DOI： 10.14188/j.1671-8844.2025.0354.

摘要

为解决单一机器学习位移监控模型泛化能力不足与可解释性弱的问题，以拱坝为例构建基于随机森林算法的位移高精度监控模型，利用SHAP(Shapley Additive Explanations)框架量化水压、温度及时效三类模型因子对顺河向位移的贡献并对总位移进行解构，以揭示因子作用机理。算例结果表明：水压因子是引起大坝顺河向位移的主要因素，其影响随库水位的抬升而非线性增加，1288m以上水位贡献超70%；1285.1m被识别为YY拱坝位移响应的临界水位，超过该阈值时水压因子对位移影响由负向转正向；模型能有效识别并剔除存在共线性的冗余温度因子，显著提升关键温度因子的贡献度；时效因子对位移贡献无偏向性。

Abstract

To address the limitations of insufficient generalization capability and weak interpretability in individual machine learning models for dam safety monitoring models

this study constructs a high-accuracy displacement monitoring model based on Random Forest algorithm

using the YY arch dam as a case study. The SHAP framework is employed to quantify the contributions of three model factors - water pressure

temperature

and time-dependent factors to the streamwise displacement. The total displacement is decomposed into three components based on feature attributes

thereby revealing the mechanism of environmental variables on dam displacement. Results demonstrate that water pressure is the primary factor

with its contribution increasing non-linearly with rising reservoir water levels

exceeding 70% at levels above 1288m. A critical water level of 1285.1m is identified for YY's displacement response

beyond which direct water pressure factors exert a positive influence on displacement. The model effectively identifies and eliminates redundant temperature factors with collinearity

significantly enhancing the contribution of key temperature factors. The time-dependent factors exhibit no directional bias in its contribution to displacement.

关键词

Keywords

references

Liu Xingyang , Li Zhanchao , Sun Linsong , et al . A critical review of statistical model of dam monitoring data [J ] . Journal of Building Engineering , 2023 , 80 : 108 - 106 .

苏荟 , 胡江 , 李星 , 等 . 改进温度分量的高海拔混凝土坝位移监控模型 [J ] . 水利水电科技进展 , 2025 , 45 ( 02 ): 106 - 112 .

SU Hui , HU Jiang , LI Xing , et al . Displacement monitoring model of high-altitude concrete dam with improved temperature component [J ] . Advances in Science and Technology of Water Resources , 2025 , 45 ( 02 ): 106 - 112 .

Yang J , Su H , Liu H , et al . A multi-point combined prediction model for deformation monitoring of concrete dams [J ] . Engineering Applications of Artificial Intelligence , 2025 , 159 ( PA ): 111652 - 111652 .

王鑫鑫 , 凌薇 , 艾学轶 . 基于变分模态分解和集成深度学习的短期风速预测 [J ] . 武汉大学学报(工学版) , 2025 , 58 ( 04 ): 677 - 686 .

WANG Xinxin , LING Wei , AI Xueyi . Short-term wind speed prediction based on variational mode decomposition and integrated deep learning [J ] . Engineering Journal of Wuhan University , 2025 , 58 ( 04 ): 677 - 686 .

苏卫星 , 冉顺义 , 刘芳 , 等 . 基于相关性变量筛选偏最小二乘回归的多维相关时间序列建模方法 [J ] . 信息与控制 , 2021 , 50 ( 04 ): 395 - 402 .

SU Weixing , RAN Shunyi , LIU Fang , et al . Modeling method of multidimensional correlation time series based on correlation variable selection partial least squares regression [J ] . Information and Control , 2021 , 50 ( 04 ): 395 - 402 .

陈立秋 , 顾冲时 , 邵晨飞 , 等 . 基于相似测点对比的大坝变形监测数据粗差识别方法 [J ] . 水利水电科技进展 , 2024 , 44 ( 04 ): 72 - 77 .

CHEN Liqiu , GU Chongshi , SHAO Chenfei , et al . Gross error identification method for dam deformation monitoring data based on similar measurement point comparison [J ] . Advances in Science and Technology of Water Resources , 2024 , 44 ( 04 ): 72 - 77 .

B K , Wu Z Y . Data-based model with EMD and a new model selection criterion for dam health monitoring [J ] . Engineering Structures , 2022 , 260 : 114171 .

程小龙 , 张斌 , 刘相杰 , 等 . 基于CEEMDAN-GMDH-ARIMA的大坝变形预测模型研究 [J ] . 人民黄河 , 2024 , 46 ( 01 ): 146 - 150 .

CHENG Xiaolong , ZHANG Bin , LIU Xiangjie , et al . Research on dam deformation prediction model based on CEEMDAN-GMDH-ARIMA [J ] . Yellow River , 2024 , 46 ( 01 ): 146 - 150 .

黄海燕 , 艾星星 , 刘兴阳 , 等 . 基于可解释性分析的大坝变形监控模型对比研究 [J ] . 人民长江 , 2024 , 55 ( 09 ): 203 - 209 .

HUANG Haiyan , AI Xingxing , LIU Xingyang , et al . Comparison of monitoring model for dam deformation based on interpretability analysis [J ] . Yangtze River , 2024 , 55 ( 09 ): 203 - 209 .

谷宇 , 苏怀智 , 张帅 , 等 . 基于自适应时序分解筛选的大坝变形预测模型 [J ] . 水利学报 , 2024 , 55 ( 09 ): 1045 - 1057+1070 .

GU Yu , SU Huaizhi , ZHANG Shuai , et al . Dam deformation prediction model based on self-adaptive temporal decomposition screening [J ] . Journal of Hydraulic Engineering , 2024 , 55 ( 09 ): 1045 - 1057+1070 .

柳聪聪 , 张锋 , 胡超 , 等 . 基于可解释性因子选择的多模型耦合式大坝变形预测方法 [J/OL ] . 长江科学院院报 , 1 - 13 [ 2025-09-26 ] . http://kns.cnki.net/kcms/detail/42.1171.TV.20250124.1454.032.html http://kns.cnki.net/kcms/detail/42.1171.TV.20250124.1454.032.html .

LIU Congcong , ZHANG Feng , HU Chao , et al . A multi-model coupled dam deformation prediction method based on interpretable factor selection [J/OL ] . Journal of Yangtze River Scientific Research Institute , 1 - 13 [ 2025-09-26 ] . http://kns.cnki.net/kcms/detail/42.1171.TV.20250124.1454.032.html http://kns.cnki.net/kcms/detail/42.1171.TV.20250124.1454.032.html .

余红玲 , 王晓玲 , 任炳昱 , 等 . 土石坝渗流性态分析的IAO-XGBoost集成学习模型与预测结果解释 [J ] . 水利学报 , 2023 , 54 ( 10 ): 1195 - 1209 .

YU Hongling , WANG Xiaoling , REN Bingyu , et al . IAO-XGBoost ensemble learning model for seepage behavior analysis of earth-rock dam and interpretation of prediction results [J ] . Journal of Hydraulic Engineering , 2023 , 54 ( 10 ): 1195 - 1209 .

马春辉 , 余飞 , 程琳 , 等 . 特高拱坝变形机理可解释性智能预测模型 [J ] . 水力发电学报 , 2024 , 43 ( 10 ): 107 - 120 .

MA Chunhui , YU Fei , CHENG Lin , et al . Explanatory intelligent prediction model for deformation mechanism of super-high arch dam [J ] . Journal of Hydroelectric Engineering , 2024 , 43 ( 10 ): 107 - 120 .

LUNDBERG S M , LEE S I . A unified approach to interpreting model predictions [J ] . Arxiv Preprint Arxiv: 2017 , 1705 : 078 - 74 .

向鑫 , 郭生练 , 崔震 , 等 . 考虑降水时空分布的洪水预报深度学习模型研究 [J ] . 武汉大学学报(工学版) , 2025 , 58 ( 06 ): 845 - 855 .

XIANG Xin , GUO Shenglian , CUI Zheng , et al . Deep learning model for flood forecasting considering the spatio-temporal distribution of precipitation [J ] . Engineering Journal of Wuhan University , 2025 , 58 ( 06 ): 845 - 855 .

田菊飞 , 苏怀智 . 基于随机森林算法的大坝应力预测模型的构建及其应用 [J ] . 水电能源科学 , 2018 , 36 ( 5 ): 54 - 56,5 .

TIAN Jufei , SU Huaizhi . Development and application of dam stress prediction model based on random forest algorithm [J ] . Water Resources and Power , 2018 , 36 ( 5 ): 54 - 56,5 .

Li K , Xu W Y , Wang Y , et al . Enhancing fuel cell system fault diagnosis:Using a bayesian optimized convolutional neural network-random forest hybrid model [J ] . Sustainable Energy Technologies and Assessments , 2025 , 82104479 - 104479 .

黄佳健 , 王宝华 , 韩杰 . 基于LightGBM和可解释代理模型的风电场并网振荡稳定分析与控制 [J ] . 武汉大学学报(工学版) , 2024 , 57 ( 10 ): 1424 - 1433 .

HUANG Jiajian , WANG Baohua , HAN Jie . Oscillation stability analysis and control of grid-connected wind farms based on LightGBM and interpretable surrogate model [J ] . Engineering Journal of Wuhan University , 2024 , 57 ( 10 ): 1424 - 1433 .

黄耀英 , 何一洋 , 沈振中 , 等 . 大坝监测量最佳统计模型优选方法 [J ] . 水利学报 , 2022 , 53 ( 02 ): 154 - 164 .

HUANG Yaoying , HE Yiyang , SHEN Zhenzhong , et al . Optimization method of optimal statistical model of dam monitoring data [J ] . Journal of Hydraulic Engineering , 2022 , 53 ( 02 ): 154 - 164 .

吴中如 . 水工建筑物安全监控理论及其应用 [M ] . 北京 : 高等教育出版社 , 2003 .

WU Zhongru . Theory and Application of Safety Monitoring for Hydraulic Structures [M ] . Beijing : Higher Education Press , 2003 .

中华人民共和国水利部 . SL601-2013 混凝土坝安全监测技术规范 [S ] . 北京 : 奔流电子音像出版（北京）有限公司 , 2013 .

Ministry of Water Resources of the People's Republic of China . SL601-2013 Technical Specification for Safety Monitoring of Concrete Dams [S ] . Beijing : Torrent Electronic & Audio-Video Publishing(Beijing)Co.,Ltd. , 2013 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Machine learning based optimization study of charging pile ventilation grille structure

Water demand forecasting based on interpretable machine learning models and Monte Carlo method

Digital evaluation of the transmission tower operation status based on stress monitoring

Research on an automatic measurement model for crop growth parameters based on machine vision

A Bayesian ensemble probabilistic prediction method for the debris⁃flow runout distance based on fusion of multiple machine learning algorithms

Related Author

CHENG Han

ZHANG Nan

YUE Yanan

WANG Huijiu

YANG Limin

WU Shijing

WU Fanyu

FU Xiang

Related Institution

Bull Industrial Research Institute of New Energy， Wuhan University

School of Power and Mechanical Engineering， Wuhan University

New Energy Division， Bull Group

Hubei Water Saving Research Center， Hubei Water Resources Research Institute

Hubei Water Resources and Hydropower Planning， Survey and Design Institute Co.， Ltd.

AI问答

⁰