Peak Breach Discharge Prediction for Landslide Dams Using a Bayesian-Optimized XGBoost Model and Sensitivity Analysis
-
摘要:
准确且迅速地评估堰塞坝溃决洪峰流量，对应急抢险至关重要.基于机器学习方法预测突发型堰塞坝溃决参数是当前的研究热点，而目前堰塞坝数据库缺少足够案例量，且堰塞坝溃决洪峰流量预测模型无法考虑各影响因素之间的非线性映射关系，这导致模型的泛化能力弱.基于此，采用泥沙冲刷模型模拟堰塞坝溃决过程，从而扩充堰塞坝溃决案例数据库；建立贝叶斯算法优化的极端梯度提升（XGBoost）的机器学习算法；提出考虑堰塞坝几何形态参数（坝高、坝宽、坝长、坝体积）、堰塞湖库容、诱发因素、物质组成（侵蚀度和结构类型）等8个影响因素的非均质堰塞坝溃决洪峰流量机器学习预测模型；基于参数敏感性分析，进一步建立便于堰塞湖灾害应急抢险使用的简化三参数模型.结果表明，与传统模型相比，贝叶斯优化的XGBoost模型具有更高的预测精度；基于唐家山和白格堰塞坝案例分析证实本文模型预测溃决洪峰流量与真实值最大误差约20%.研究成果能够为堰塞坝应急抢险地质处置及区域防灾减灾提供有益参考.
Abstract:Accurate and rapid assessment of peak flood discharge from landslide dam breaches is crucial for emergency response efforts. Predicting breach parameters of sudden landslide dam failures using machine learning methods has become a current research focus. However, existing landslide dam database slack sufficient case records, and current predictive models for peak breach discharge fail to capture the nonlinear interactions among influencing factors, resulting in limited generalization capability. In response to this, this study employs a sediment erosion model to simulate the landslide dam breach process, thereby expanding the landslide dam breach case database. The Extreme Gradient Boosting (XGBoost) machine learning algorithm is used to predict the peak flood discharge of landslide dam breaches, and the Bayesian algorithm is employed to optimize the hyperparameters of the XGBoost model. An innovative machine learning prediction model for peak flood discharge of heterogeneous landslide dam breaches is proposed, considering eight influencing factors, including geometric parameters of the dam (height, width, length, volume), reservoir capacity, triggering factors, and material composition (erodibility and structural type). The results indicate that, compared to traditional models, the Bayesian-optimized XGBoost machine learning model exhibits higher prediction accuracy. Case analyses of Tangjiashan and Baige landslide dams confirm that the model's predicted peak flood discharge has a maximum error of approximately 20% compared to the actual values. This study provides a valuable reference for emergency response and regional disaster mitigation in the context of landslide dam breaches.
-
Key words:
- landslide dam /
- dam breach /
- peak flood discharge /
- Bayesian optimization /
- XGBoost /
- engineering geology
-
图 1 基于贝叶斯优化XGBoost的堰塞坝溃决洪峰流量预测模型构建方法
Fig. 1. Method for constructing the prediction model of peak flood discharge of landslide dam breach based on Bayesian Optimization XGBoost
图 2 贝叶斯优化XGBoost预测模型全数据(a)和测试集(b)预测结果
Fig. 2. Prediction results on full data set (a) and test data set (b)
图 3 预测模型在测试集中泛化能力对比
Fig. 3. Comparison of generalization ability of prediction model in test set
图 5 三参数预测模型全数据集预测结果(a)和测试集预测结果(b)
Fig. 5. Three-parameter prediction model results in full data (a) and test data (b)
表 1 堰塞坝真实形态参数区间
Table 1. Parameter interval of real landslide dams
坝体类型 坝高∶坝顶宽 坝高∶坝底宽 上游坝坡坡度 下游坝坡坡度 坝体模型 1 0.2 26.56° 26.56° 堰塞坝案例库 0.2~3 0.02~1 11°~45° 11°~45° 
下载: 导出CSV
表 2 模型优化参数
Table 2. Optimized model parameters
超参数名称 含义 范围 最终值 n_ estimators 决策树数量 （20, 100） 40 learning_ rate 学习率 （0.01, 1） 0.03 max_ depth 最大树深度 （1, 10） 6
下载: 导出CSV
表 3 不同堰塞坝溃决洪峰流量预测模型
Table 3. Different forecasting models for breach peak discharge of landslide dams
模型 参考文献 案例数 模型特征 $ {Q}_{\mathrm{p}}=6.3{H}_{\mathrm{d}}^{1.59} $ Costa and Schuster (1988) 36 单参数简单 $ {Q}_{\mathrm{p}}=3.130{H}_{\mathrm{d}}^{0.120}{W}_{\mathrm{d}}^{0.302}{V}_{\mathrm{d}}^{-0.106}{V}_{\mathrm{l}}^{0.453}{e}^{a} $ 石振明等(2014) 41 考虑坝体侵蚀度 $ \frac{{Q}_{\mathrm{p}}}{{g}^{1/2}{H}_{\mathrm{d}}^{5/2}}=0.828{\left(\frac{{H}_{\mathrm{d}}}{{H}_{\mathrm{r}}}\right)}^{-0.128}{\left(\frac{{H}_{\mathrm{d}}}{{W}_{\mathrm{d}}}\right)}^{-0.432}{\left(\frac{{V}_{\mathrm{d}}^{1/3}}{{H}_{\mathrm{d}}}\right)}^{-0.394}{\left(\frac{{V}_{\mathrm{l}}^{1/3}}{{H}_{\mathrm{d}}}\right)}^{1.151} $ 齐子杰等(2022) 57 未考虑坝体侵蚀度 $ \frac{{Q}_{\mathrm{p}}}{{g}^{1/2}{H}_{\mathrm{d}}^{5/2}}={\left(\frac{{H}_{\mathrm{d}}}{{H}_{\mathrm{r}}}\right)}^{-1.417}{\left(\frac{{H}_{\mathrm{d}}}{{W}_{\mathrm{d}}}\right)}^{-0.265}{\left(\frac{{V}_{\mathrm{d}}^{1/3}}{{H}_{\mathrm{d}}}\right)}^{-0.471}{\left(\frac{{V}_{\mathrm{l}}^{1/3}}{{H}_{\mathrm{d}}}\right)}^{1.569}{e}^{a} $ Peng and Zhang (2012) 45 考虑坝体侵蚀度的多参数模型 贝叶斯优化XGBoost机器学习模型 本文模型 122 考虑坝体侵蚀度的多参数模型 注：$ {Q}_{\mathrm{p}} $溃决洪峰流量，$ {H}_{\mathrm{d}} $堰塞坝高度，$ {W}_{\mathrm{d}} $堰塞坝宽度，$ {V}_{\mathrm{d}} $堰塞坝体积量，$ {V}_{\mathrm{l}} $堰塞湖库容，a堰塞坝体侵蚀度.
下载: 导出CSV
表 5 堰塞坝参数
Table 5. Landslide dam parameters
名称 坝高(m) 坝宽(m) 坝长(m) 坝体体积(m3) 库容(m3) 侵蚀度 结构类型 诱因 唐家山 82 611.8 406 20.4×106 247×106 中等 块石夹土 地震 白格 61 1 400 580 26×106 290×106 高 土夹碎石 山体失稳
下载: 导出CSV
表 6 不同堰塞坝溃决洪峰流量预测模型计算结果对比
Table 6. Comparison of calculation results of prediction models of breach peak discharge
案例名称 实测值(m3·s-1) 结果对比 本文全参数模型 本文三参数模型 唐家山 6 500 预测值(m3·s-1) 7 242.3 7 661.55 相对误差(%) 11.42 17.87 白格 10 000 预测值(m3·s-1) 11 013 12 175 相对误差(%) 10.13 21.75
下载: 导出CSV
-
Bergen, K. J., Johnson, P. A., de Hoop, M. V., et al., 2019. Machine Learning for Data-Driven Discovery in Solid Earth Geoscience. Science, 363(6433): eaau0323. https://doi.org/10.1126/science.aau0323 Cai, Y. J., Cheng, H. Y., Wu, S. F., et al., 2020. Breaches of the Baige Barrier Lake: Emergency Response and Dam Breach Flood. Science China Technological Sciences, 63(7): 1164-1176. https://doi.org/10.1007/s11431-019-1475-y Chen, X. Q., Cui, P., Cheng, Z. L., et al., 2008. Emergency Risk Assessment of Dammed Lakes Caused by the Wenchuan Earthquake on May 12, 2008. Earth Science Frontiers, 15(4): 244-249 (in Chinese with English abstract). Cheng, Z. L., Xia, Y. F., Xu, H., et al., 2024. Study on the Response Mechanism of Landslide Dam Failure to External Dynamic Condition. Advances in Water Science, 35(4): 657-668 (in Chinese with English abstract). Costa, J. E., Schuster, R. L., 1988. The Formation and Failure of Natural Dams. Geological Society of America Bulletin, 100(7): 1054-1068. https://doi.org/10.1130/0016-7606(1988)1001054:tfafon>2.3.co;2 doi: 10.1130/0016-7606(1988)1001054:tfafon>2.3.co;2 Cui, P., Zhu, Y. Y., Han, Y. S., et al., 2009. The 12 May Wenchuan Earthquake-Induced Landslide Lakes: Distribution and Preliminary Risk Evaluation. Landslides, 6(3): 209-223. https://doi.org/10.1007/s10346-009-0160-9 Feng, Z. Y., Yang, X. G., Zhou, J. W., et al., 2024. Estimation of Breach Parameters and Longevity of Landslide Dam. Journal of Hydraulic Engineering, 55(3): 367-377 (in Chinese with English abstract). Hakimzadeh, H., Nourani, V., Amini, A. B., 2014. Genetic Programming Simulation of Dam Breach Hydrograph and Peak Outflow Discharge. Journal of Hydrologic Engineering, 19(4): 757-768. https://doi.org/10.1061/(asce)he.1943-5584.0000849 Li, D. Y., Nian, T. K., Tiong, R. L. K., et al., 2023. River Blockage and Impulse Wave Evolution of the Baige Landslide in October 2018: Insights from Coupled DEM-CFD Analyses. Engineering Geology, 321: 107169. https://doi.org/10.1016/j.enggeo.2023.107169 Long, X. Y., Hu, Y. X., Gan, B. R., et al., 2024. Numerical Simulation of the Mass Movement Process of the 2018 Sedongpu Glacial Debris Flow by Using the Fluid-Solid Coupling Method. Journal of Earth Science, 35(2): 583-596. https://doi.org/10.1007/s12583-022-1625-1 Mastbergen, D. R., Van Den Berg, J. H., 2003. Breaching in Fine Sands and the Generation of Sustained Turbidity Currents in Submarine Canyons. Sedimentology, 50(4): 625-637. https://doi.org/10.1046/j.1365-3091.2003.00554.x Mei, S. Y., Chen, S. S., Zhong, Q. M., et al., 2022. Detailed Numerical Modeling for Breach Hydrograph and Morphology Evolution during Landslide Dam Breaching. Landslides, 19(12): 2925-2949. https://doi.org/10.1007/s10346-022-01952-1 Nian, T. K., Wu, H., Chen, G. Q., et al., 2018. Research Progress on Stability Evaluation Method and Disaster Chain Effect of Landslide Dam. Chinese Journal of Rock Mechanics and Engineering, 37(8): 1796-1812 (in Chinese with English abstract). Peng, M., Zhang, L. M., 2012. Breaching Parameters of Landslide Dams. Landslides, 9(1): 13-31. https://doi.org/10.1007/s10346-011-0271-y Qi, Z. J., Huang, W., Wang, L. X., et al., 2022. Evaluation and Optimization Analysis of Barrier Dam Breach Parameter Models. Yangtze River, 53(12): 157-166 (in Chinese with English abstract). Reichstein, M., Camps-Valls, G., Stevens, B., et al., 2019. Deep Learning and Process Understanding for Data-Driven Earth System Science. Nature, 566(7743): 195-204. https://doi.org/10.1038/s41586-019-0912-1 Ruan, H. C., Chen, H. Y., Li, X., et al., 2022. Modification and Comparison of Flood Peak Discharge Prediction Model for Landslide Dam Failure Based on New Samples. Science Technology and Engineering, 22(20): 8606-8615 (in Chinese with English abstract). Shan, Y. B., Chen, S. S., Zhong, Q. M., 2020. Rapid Prediction of Landslide Dam Stability Using the Logistic Regression Method. Landslides, 17(12): 2931-2956. https://doi.org/10.1007/s10346-020-01414-6 Shen, D. Y., Shi, Z. M., Zheng, H. C., et al., 2022. Effects of Grain Composition on the Stability, Breach Process, and Breach Parameters of Landslide Dams. Geomorphology, 413: 108362. https://doi.org/10.1016/j.geomorph.2022.108362 Shi, N., Li, Y. L., Wen, L. F., et al., 2022. Rapid Prediction of Landslide Dam Stability Considering the Missing Data Using XGBoost Algorithm. Landslides, 19(12): 2951-2963. https://doi.org/10.1007/s10346-022-01947-y Shi, Z. M., Ma, X. L., Peng, M., et al., 2014. Statistical Analysis and Efficient Dam Burst Modelling of Landslide Dams Based on a Large-Scale Database. Chinese Journal of Rock Mechanics and Engineering, 33(9): 1780-1790 (in Chinese with English abstract). Soulsby, R. L., 1997. Dynamics of Marine Sands: A Manual for Practical Applications. Oceanographic Literature Review, 9: 947. Walder, J. S., O'Connor, J. E., 1997. Methods for Predicting Peak Discharge of Floods Caused by Failure of Natural and Constructed Earthen Dams. Water Resources Research, 33(10): 2337-2348. https://doi.org/10.1029/97WR01616 Wu, H., 2021. Study on the Process Simulation and Hazard Prediction Method of Landslide Damming (Dissertation). Dalian University of Technology, Dalian (in Chinese with English abstract). Wu, H., Nian, T. K., Shan, Z. G., 2023. Investigation of Landslide Dam Life Span Using Prediction Models Based on Multiple Machine Learning Algorithms. Geomatics, Natural Hazards and Risk, 14(1): 2273213. https://doi.org/10.1080/19475705.2023.2273213 Wu, H., Nian, T. K., Shan, Z. G., 2023. Research Progress on Formation and Evolution Mechanism and Risk Prediction Method of Landslide Blocking River and Dam. Chinese Journal of Rock Mechanics and Engineering, 42(S1): 3192-3205 (in Chinese with English abstract). Yuan, H., Guo, C. B., Wu, R. A., et al., 2023. Research Progress and Prospects of the Giant Yigong Long Run-out Landslide, Tibetan Plateau, China. Geological Bulletin of China, 42(10): 1757-1773 (in Chinese with English abstract). Zhang, R. H., Li, Y. Q., Goh, A. T. C., et al., 2021. Analysis of Ground Surface Settlement in Anisotropic Clays Using Extreme Gradient Boosting and Random Forest Regression Models. Journal of Rock Mechanics and Geotechnical Engineering, 13(6): 1478-1484. https://doi.org/10.1016/j.jrmge.2021.08.001 Zhang, W. G., Wu, C. Z., Tang, L. B., et al., 2023. Efficient Time-Variant Reliability Analysis of Bazimen Landslide in the Three Gorges Reservoir Area Using XGBoost and LightGBM Algorithms. Gondwana Research, 123: 41-53. https://doi.org/10.1016/j.gr.2022.10.004 Zhu, X. H., Liu, B. X., Guo, J., et al., 2020. Summary of Research on Landslide Dam Break. Science Technology and Engineering, 20(21): 8440-8451 (in Chinese with English abstract). 陈晓清, 崔鹏, 程尊兰, 等, 2008.5. 12汶川地震堰塞湖危险性应急评估. 地学前缘, 15(4): 244-249. 成泽霖, 夏云峰, 徐华, 等, 2024. 滑坡堰塞坝漫顶溃决对外部动力条件的响应机制. 水科学进展, 35(4): 657-668. 冯震宇, 杨兴国, 周家文, 等, 2024. 堰塞坝溃决参数及其寿命预测. 水利学报, 55(3): 367-377. 年廷凯, 吴昊, 陈光齐, 等, 2018. 堰塞坝稳定性评价方法及灾害链效应研究进展. 岩石力学与工程学报, 37(8): 1796-1812. 齐子杰, 黄卫, 汪利先, 等, 2022. 堰塞坝溃决参数模型评估与优化分析. 人民长江, 53(12): 157-166. 阮合春, 陈华勇, 李霄, 等, 2022. 基于新样本的滑坡坝溃决洪峰流量预测模型修正与对比. 科学技术与工程, 22(20): 8606-8615. 石振明, 马小龙, 彭铭, 等, 2014. 基于大型数据库的堰塞坝特征统计分析与溃决参数快速评估模型. 岩石力学与工程学报, 33(9): 1780-1790. 吴昊, 2021. 滑坡堵江成坝过程模拟及危险性预测方法研究(博士学位论文). 大连: 大连理工大学. 吴昊, 年廷凯, 单治钢, 2023. 滑坡堵江成坝的形成演进机制及危险性预测方法研究进展. 岩石力学与工程学报, 42(增刊1): 3192-3205. 袁浩, 郭长宝, 吴瑞安, 等, 2023. 西藏易贡高位远程滑坡研究进展与展望. 地质通报, 42(10): 1757-1773. 朱兴华, 刘邦晓, 郭剑, 等, 2020. 堰塞坝溃坝研究综述. 科学技术与工程, 20(21): 8440-8451. -
点击查看大图
计量
- 文章访问数: 257
- HTML全文浏览量: 21
- PDF下载量: 43
- 被引次数: 0
