Research on Physics-Guided Intelligent Prediction Model for Valley Contraction Deformation of Reservoir Bank Slopes
-
摘要: 库岸谷幅持续收缩变形会威胁大坝安全，因此谷幅变形预测对大坝安全管理至关重要.以溪洛渡水电站谷幅变形为研究对象，提出一种基于物理引导的智能预测模型，开展谷幅变形预测. 基于谷幅变形规律分析，建立统计回归物理模型（STPM），获取不同诱发因素对谷幅变形贡献的位移分量. 据此，以各分量为输入，联合灰色关联度分析（GRA）、长短期记忆网络（LSTM）和随机森林（RF），构建基于物理引导的机器学习预测框架（STPM-GRA-LSTM-RF），开展不确定性预测分析. 研究结果表明：（1）溪洛渡谷幅变形主要由黏塑性变形、黏弹性变形和有效应力引起，对总位移平均贡献率分别为67.71%、29.75%及2.51%；（2）与STPM模型、LSTM模型、SVM模型和XGBoost模型相比，本文提出模型的预测精度更高，并因其考虑了谷幅变形机制，可靠性显著提升. 研究成果为类似库岸高边坡的变形预测与安全管控提供有价值的参考.Abstract: Reservoir impoundment-induced valley contraction deformation poses a significant threat to the overall safety of dams, making the prediction of valley width deformation crucial for dam safety management. This study focuses on the valley deformation characteristics of the Xiluodu Hydropower Station and proposes a physics-guided intelligent prediction model to conduct valley deformation forecasting. Based on the analysis of valley deformation patterns, a statistical regression physics-based model (STPM) is developed to quantify the displacement components contributed by individual triggering factors to valley deformation. Building upon this foundation, a machine learning prediction framework (STPM-GRA-LSTM-RF) integrating grey relational analysis (GRA), long short-term memory networks (LSTM), and random forest algorithms (RF) is constructed with displacement components as inputs, incorporating uncertainty analysis in valley deformation prediction. Key findings include: (1) The valley deformation at Xiluodu is primarily caused by viscoelastic deformation, viscoplastic deformation, and effective stress-induced deformation, contributing 67.71%, 29.75%, and 2.51%, respectively, to the total displacement; (2) Compared with the STPM, LSTM, SVM, and XGBoost models, the proposed STPM-GRA-LSTM-RF model delivers higher predictive accuracy, moreover, by incorporating the physico-mechanical mechanisms of valley deformation, it significantly enhances the reliability of the predictions. This research provides valuable insights for deformation prediction and safety control of high reservoir slopes in analogous projects.
-
图 1 溪洛渡水库鸟瞰图（来源于Google Earth）
Fig. 1. Aerial view of Xiluodu Reservoir (from Google Earth)
图 3 溪洛渡水库谷幅测线布置平面图
Fig. 3. Configuration of valley measuring lines for Xiluodu Reservoir
图 4 溪洛渡谷幅变形与库水位监测曲线
Fig. 4. Monitoring curves of valley deformation and reservoir water level at Xiluodu Reservoir
图 5 VD3谷幅变形、变形速率及库水位曲线
Fig. 5. Valley deformation, deformation velocity and reservoir water level of the measuring line VD3
图 6 不同库水条件下VD3谷幅变形时间曲线
Fig. 6. Displacement-time curve of the measuring line VD3 under different reservoir water conditions
图 8 基于物理引导的谷幅变形预测流程图
Fig. 8. Flowchart of physics-guided prediction model for valley deformation
图 12 灰色关联度计算结果
a.VD1灰色关联度计算结果；b. VD3灰色关联度计算结果；c. VD5灰色关联度计算结果
Fig. 12. Calculation results of grey correlation degree
图 14 不同模型预测区间对比
Fig. 14. Comparative analysis of prediction intervals across different models
图 17 模型外推预测结果
a. VD1；b. VD3；c. VD5
Fig. 17. Extrapolated predictive outputs of the model
表 1 统计回归物理模型参数计算结果
Table 1. Calculation results of model parameters in statistical regression physical model
谷幅测线 α1 α2 α3 b1 α4 R2 VD1 0.050 105.649 2.336 23.898 0.015 0.97 VD3 0.067 91.995 2.930 21.885 0.015 0.97 VD5 0.046 113.745 2.636 29.000 0.012 0.96 
下载: 导出CSV
表 2 超参数设置表
Table 2. Configuration settings of hyperparameters
参数 定义 设置值 batch_size 每次训练选取的数据样本数量 64 optimizer 优化调整网络的权重和偏置 Adam epoch 模型训练时期次数 300~400 dropout 模型训练阶段忽略单元 0.01 hidden_size 隐藏层的维度大小 32、128 损失函数 衡量预测值与真实值的差距 Huber 隐藏层数 神经网络中的中间层 2
下载: 导出CSV
表 3 点预测评价指标对比
Table 3. Comparison of point prediction evaluation metrics
测线 MAE MSE RMSE MAPE R2 VD1a 0.99 1.49 1.22 2.24% 0.99 VD1b 2.01 6.35 1.49 5.22% 0.97 VD1c 2.92 14.25 3.78 6.81% 0.94 VD1d 1.43 3.89 1.97 4.83% 0.98 VD1e 2.02 6.24 2.50 6.57% 0.97 VD1d 2.02 6.24 2.50 6.57% 0.97 VD3a 1.18 1.95 1.40 2.22% 0.99 VD3b 2.02 7.32 2.71 4.93% 0.98 VD3c 3.36 18.51 4.30 6.95% 0.94 VD3d 1.67 5.01 2.23 3.61% 0.98 VD3e 2.78 11.54 3.40 7.35% 0.97 VD5a 0.92 1.38 1.17 1.89% 0.99 VD5b 1.72 6.07 2.46 4.51% 0.97 VD5c 3.05 15.55 3.94 6.63% 0.93 VD5d 1.57 4.34 2.08 3.54% 0.98 VD5e 2.77 9.99 3.16 5.11% 0.96 注：a代表STPM-GRA-LSTM-RF；b代表LSTM模型；c代表SVM模型；d代表XGBoost模型；e代表STPM模型
下载: 导出CSV
-
Barla, G., Paronuzzi, P., 2013. The 1963 Vajont Landslide: 50th Anniversary. Rock Mechanics and Rock Engineering, 46(6): 1267-1270. https://doi.org/10.1007/s00603-013-0483-7 Chang, X. Y., Guo, J., Qin, H., et al., 2024. Single-Objective and Multi-Objective Flood Interval Forecasting Considering Interval Fitting Coefficients. Water Resources Management, 38(10): 3953-3972. https://doi.org/10.1007/s11269-024-03848-2 Chen, M. X., Feng, A. J., Wei, W., et al., 2024. Statistical Analysis of Long-Term Deformations and Determination of Warning Thresholds for Near-Dam Reservoir Bank Slopes. Bulletin of Engineering Geology and the Environment, 83(11): 437. https://doi.org/10.1007/s10064-024-03928-y Chen, M. X., Tao, P. J., Zhou, C. B., et al., 2024. Model Modification of Verhulst Inverse-Function Forecasting Model and Probabilistic Forecast for Landslide Failure Time. Earth Science, 49 (5): 1692-1705(in Chinese with English abstract). Chen, M. X., Wei, W., Jiang, Q. H., 2022. Use of Quantile Regression with Fukui-Okubo Model for Prediction and Early Warning of Reservoir Bank Slope Failure. Rock Mechanics and Rock Engineering, 55(11): 7145-7169. https://doi.org/10.1007/s00603-022-02975-9 Devore, J. L., 2000. Probability and Statistics. Pacific Grove, Brooks/Cole. Duffaut, P., 2013. The Traps Behind the Failure of Malpasset Arch Dam, France, in 1959. Journal of Rock Mechanics and Geotechnical Engineering, 5(5): 335-341. https://doi.org/10.1016/j.jrmge.2013.07.004 Gong, W. P., Zhang, S. Y., Juang, C. H., et al., 2024. Displacement Prediction of Landslides at Slope-Scale: Review of Physics-Based and Data-Driven Approaches. Earth-Science Reviews, 258: 104948. https://doi.org/10.1016/j.earscirev.2024.104948 He, Z., Liu, Y. R., Yang, Q., et al., 2018. Mechanism of ValleyDeformation of Xiluodu Arch Dam and Back Analysis and Long-Term Stability Analysis. Chinese Journal of Rock Mechanics and Engineering, 37(S2): 4198-4206(in Chinese with English abstract). Huang, F. M., Ou, Y. W. P., Jiang, S. H., et al., 2024. Landslide Susceptibility Prediction Considering Spatio-Temporal Division Principle of Training/Testing Datasets in Machine Learning Models. Earth Science, 49(5): 1607-1618(in Chinese with English abstract). Lei, Z. Q., 2018. DDA Analysis on Creep Deformation of SlopeTriggered by Reservoir Impoundment(Dissertation). China Institute of Water Resources and Hydropower, Beijing(in Chinese with English abstract). Li, B. Y., Wang, G. L., Chen, L. C., et al., 2024. Analysis of Landslide Deformation Mechanisms and Coupling Effects under Rainfall and Reservoir Water Level Effects. Engineering Geology, 343: 107803. https://doi.org/10.1016/j.enggeo.2024.107803 Li, M. W., Selvadurai, A. P. S., Zhou, Z. F., 2023. Observations and Computational Simulation of River Valley Contraction at the Xiluodu Dam, Yunnan, China. Rock Mechanics and Rock Engineering, 56(6): 4109-4131. https://doi.org/10.1007/s00603-023-03269-4 Li, X. Y., Zhang, L. M., Jiang, S. H., 2016. Updating Performance of High Rock Slopes by Combining Incremental Time-Series Monitoring Data and Three-Dimensional Numerical Analysis. International Journal of Rock Mechanics and Mining Sciences, 83: 252-261. https://doi.org/10.1016/j.ijrmms.2014.09.011 Ma, J. W., Tang, H. M., Liu, X., et al., 2018. Probabilistic Forecasting of Landslide Displacement Accounting for Epistemic Uncertainty: a Case Study in the Three Gorges Reservoir Area, China. Landslides, 15(6): 1145-1153. https://doi.org/10.1007/s10346-017-0941-5 Reyad, M., Sarhan, A. M., Arafa, M., 2023. A Modified Adam Algorithm for Deep Neural Network Optimization. Neural Computing and Applications, 35(23): 17095-17112. https://doi.org/10.1007/s00521-023-08568-z Rong, G., Xu, L. D., Yu, S., et al., 2024. The Mechanism of Large-Scale River Valley Deformation Induced by Impoundment at the Baihetan Hydropower Station. Computers and Geotechnics, 165: 105898. https://doi.org/10.1016/j.compgeo.2023.105898 Sun, Y., Huang, J. S., Jin, W., et al., 2019. Bayesian Updating for Progressive Excavation of High Rock Slopes Using Multi-Type Monitoring Data. Engineering Geology, 252: 1-13. https://doi.org/10.1016/j.enggeo.2019.02.013 Sun, M. C., Xu, W. Y., Wang, H. L., et al., 2020. A Novel Hybrid Intelligent Prediction Model for Valley Deformation: a Case Study in Xiluodu Reservoir Region, China. Computers, Materials & Continua, 66(1): 1057-1074. https://doi.org/10.32604/cmc.2020.012537 Sun, Z. G., Wang, G. T., Li, P. F., et al., 2024. An Improved Random Forest Based on the Classification Accuracy and Correlation Measurement of Decision Trees. Expert Systems with Applications, 237: 121549. https://doi.org/10.1016/j.eswa.2023.121549 Wang, D. Y., Zhu, H. H., Wu, B., et al., 2024. Performance Evaluation of Underground Pipelines Subjected to Landslide Thrust with Fiber Optic Strain Sensing Nerves. Acta Geotechnica, 19(10): 6993-7009. https://doi.org/10.1007/s11440-024-02311-1 Wang, L. Q., Xiao, T., Liu, S. L., et al., 2023. Quantification of Model Uncertainty and Variability for Landslide Displacement Prediction Based on Monte Carlo Simulation. Gondwana Research, 123: 27-40. https://doi.org/10.1016/j.gr.2023.03.006 Wang, N. T., Liu, L., Shi, T. T., et al., 2023. Study of the Impact of Reservoir Water Level Decline on the Stability Treated Landslide on Reservoir Bank. Alexandria Engineering Journal, 65: 481-492. https://doi.org/10.1016/j.aej.2022.10.042 Wang, S. G., Liu, Y. R., Yang, Q., et al., 2020. Analysis of the Abutment Movements of High Arch Dams Due to Reservoir Impoundment. Rock Mechanics and Rock Engineering, 53(5): 2313-2326. https://doi.org/10.1007/s00603-020-02059-6 Wang, Y., Yang, Q., Zhang, M., et al., 2023. Study on Mechanism of Reservoir Valley Width Deformation During Impoundment of Xiluodu Arch Dam. Chinese Journal of Rock Mechanics and Engineering, 42(5): 1083-1095(in Chinese with English abstract). Wen, H. J., Liu, L., Zhang, J. L., et al., 2023. A Hybrid Machine Learning Model for Landslide-Oriented Risk Assessment of Long-Distance Pipelines. Journal of Environmental Management, 342: 118177. https://doi.org/10.1016/j.jenvman.2023.118177 Wu, P. C., Tan, D. Y., Lin, S. Q., et al., 2022. Development of a Monitoring and Warning System Based on Optical Fiber Sensing Technology for Masonry Retaining Walls and Trees. Journal of Rock Mechanics and Geotechnical Engineering, 14(4): 1064-1076. https://doi.org/10.1016/j.jrmge.2021.09.013 Yang, Q., Pan, Y. Y., Cheng, L., et al., 2015. Mechanism of Valley Deformation of High Arch Dam and Effective Stress Principle for Unsaturated Fractured Rock Mass. Chinese Journal of Rock Mechanics and Engineering, 34(11): 2258-2269(in Chinese with English abstract). Yang, Y. T., Dai, Z. W., Lu, Y. S., et al., 2024. Deformation Characteristics and Stability Changes Characteristics of Reservoir Landslides with Double-Sliding Zones. Earth Science, 49(4): 1498-1514(in Chinese with English abstract). Yin, Y. P., Huang, B. L., Zhang, Z. H., et al., 2022. GeologicalDisaster Prevention and Control in the Three Gorges Reservoir Area. Science Press, Beijing(in Chinese). Zhang, C., Bai, Q. S., Han, P. H., 2023. A Review of Water Rock Interaction in Underground Coal Mining: Problems and Analysis. Bulletin of Engineering Geology and the Environment, 82(5): 157. https://doi.org/10.1007/s10064-023-03142-2 Zhang, W. G., Gu, X., Liu, H. L., et al., 2018. Probabilistic Back Analysis of Soil Parameters and Displacement Prediction of Unsaturated Slopes Using Bayesian Updating. Rock and Soil Mechanics, 43(4): 1112-1122(in Chinese with English abstract). Zhao, S., Lin, S. Q., Tan, D. Y., et al., 2025. Smart Prediction of Rock Crack Opening Displacement from Noisy Data Recorded by Distributed Fiber Optic Sensing. Journal of Rock Mechanics and Geotechnical Engineering, 17(5): 2619-2632. https://doi.org/10.1016/j.jrmge.2024.09.009 Zhao, S., Tan, D. Y., Lin, S. Q., et al., 2023. A Deep Learning-Based Approach with Anti-Noise Ability for Identification of Rock Microcracks Using Distributed Fibre Optic Sensing Data. International Journal of Rock Mechanics and Mining Sciences, 170: 105525. https://doi.org/10.1016/j.ijrmms.2023.105525 Zhuang, C., Zhou, Z. F., Li, M. W., et al., 2019. Prediction of Valley Shrinkage Deformation in Xiluodu Hydropower Plant Based on the Hydraulic Responses of a Confined Aquifer. Chinese Journal of Geotechnical Engineering, 41(08): 1472-1480(in Chinese with English abstract). Zhou, C. B., Jiang, Q. H., Yao, C., et al., 2025. Research Progress and Challenges to Studies on Deformation and Stability of High Steep Slopes in Hydropower Projects. Journal of Hydroelectric Engineering, 44(1): 1-17(in Chinese with English abstract). Zhou, L., Liu, S. M., Zhang, K., et al., 2024. Review on Development of Valley Deformation Monitoring Technology. Express Water Resources & Hydropower Information, 45(10): 50-54(in Chinese with English abstract). Zhou, X. P., Pan, X. K., Berto, F., 2022. A State of the Art Review on Creep Damage Mechanics of Rocks. Fatigue & Fracture of Engineering Materials & Structures, 45(3): 627-652. https://doi.org/10.1111/ffe.13625 Zhu, H. H., Gao, Y. X., Chen, D. D., et al., 2024. Interfacial Behavior of Soil-Embedded Fiber Optic Cables with Micro-Anchors for Distributed Strain Sensing. Acta Geotechnica, 19(4): 1787-1798. https://doi.org/10.1007/s11440-023-01956-8 白细民, 余圣品, 叶海龙, 等, 2021. 南昌红盆深埋藏地热热储模式探讨. 江西地学新进展—江西省地质学会2021年学术年会论文集. 南昌: 江西省地质学会, 71-76. 白细民, 邓建军, 叶海龙, 等, 2025. 江西省地热能、矿泉水资源潜力调查评价成果报告. 南昌: 江西省地质局水文地质大队. 陈墨香, 汪集旸, 1994. 中国地热资源——形成特点和潜力评估. 北京: 科学出版社, 1-39. 邓建军, 段文兵, 白细民, 等, 2025. 南昌盆地地热资源赋存特征及潜力评价. 中国地质, 1-21[2025-06-26]. http://kns.cnki.net/kcms/detail/11.1167.P.20250415.1644.007.html .胡圣标, 何丽娟, 汪集旸, 2001. 中国大陆地区大地热流数据汇编(第三版). 地球物理学报, 44(5): 611-626. 胡圣标, 黄少鹏, 2015. 中国陆地大地热流. 见: 汪集旸, 著. 地热学及其应用. 北京: 科学出版社, 64-122. 胡圣标, 熊亮萍, 汪集旸, 等, 1992. 江西省首批实测大地热流数据报道. 科学通报, (19): 1791-1793. 姜光政, 高堋, 饶松, 等, 2016. 中国大陆地区大地热流数据汇编(第四版). 地球物理学报, 59(8): 2892-2910. 江晓雪, 朱传庆, 2025. 复杂热传递作用下冀中坳陷深、浅部地热资源差异分布. 地球科学, 50(4): 1485-1498. doi: 10.3799/dqkx.2024.031 刘峰, 王贵玲, 姜光政, 等, 2024. 我国陆区大地热流测量新进展与新认识. 地学前缘, 31(6): 19-30. 刘峰, 王贵玲, 张薇, 等, 2020. 江西宁都县北部大地热流特征及地热资源成因机制. 地质通报, 39(12)1883-1890. 林乐夫, 王安东, 孙占学, 等, 2017. 江西省实测地表热流值及特征. 能源研究与管理, (3): 91-94. 李学礼, 史维浚, 周文斌, 等, 1992. 江西大地热流. 地质科学, (S1): 383-385. 邱楠生, 胡圣标, 何丽娟, 2024a. 沉积盆地地热体制研究的理论与应用. 北京: 石油工业出版社, 6-9. 邱楠生, 许晴, 朱传庆, 等, 2024b. 基于居里面深度对中国东北部大地热流的研究. 地质学报, 98(3): 680-691. 宋嘉佳, 王贵玲, 邢林啸, 等. 2023. 岩石热导率校正对大地热流计算值的影响——以渤海湾盆地冀中坳陷为例. 地质论评, 69(4): 1349-1364. 孙旭东, 郭兴伟, 张训华, 等, 2023. 南黄海盆地地层生热率及岩石圈热结构. 地球科学, 48(3): 1040-1057. doi: 10.3799/dqkx.2023.044 汪集旸, 黄少鹏. 1990. 中国大陆地区大地热流数据汇编(第二版). 地震地质, 12(4): 351-363, 366. 王一波, 刘绍文, 陈超强, 等. 2024. 中国陆域大地热流数据汇编(第五版). 地球物理学报, 67(11): 4233-4265. 王兆英, 林强, 黄海, 等, 1999. 南昌地区地热地质补充调查及地热勘查靶区论证报告. 南昌: 江西省地质局水文地质大队. 王朱亭, 张超, 姜光政, 等, 2019. 雄安新区现今地温场特征及成因机制. 地球物理学报, 62(11): 4313-4322. 杨淑贞, 张文仁, 李国桦, 等, 1993. 柴达木盆地岩石热导率的饱水试验研究及热流校正. 岩石学报, 9(2): 199-204. 余如洋, 黄少鹏, 张炯, 等, 2020. 二连盆地白音查干凹陷和乌里雅斯太凹陷岩石热导率测试与分析. 岩石学报, 36(2): 621-636. 闫晓雪, 甘浩男, 刘峰, 等, 2023. 江西赣县大地热流特征与热源机制研究. 地球学报, 44(1): 107-116. 赵永信, 杨淑贞, 张文仁, 等, 1995. 岩石热导率的温压实验及分析. 地球物理学进展, 10(1): 104-113 邹开真, 庞玉茂, 陈琰, 等, 2023. 柴达木盆地英东地区大地热流及影响因素. 地球科学, 48(3): 1002-1013. -
点击查看大图
图(18) / 表(3)
计量
- 文章访问数: 152
- HTML全文浏览量: 31
- PDF下载量: 3
- 被引次数: 0
