Landslide Displacement Prediction Based on a Deep Learning Model Considering the Attention Mechanism
-
摘要: 现有的基于数据驱动的滑坡位移预测模型大多是基于时间序列数据的单点建模，不能考虑整个边坡的变形相关性和滑坡变形的全局建模.为了克服这一缺点，本研究提出了一种基于时空注意（spatial-temporal attention，STA）机制的深度学习模型，该模型将卷积神经网络（convolutional neural network，CNN）与长短时记忆（long short-term memory）神经网络相结合.通过CNN和卷积注意力模块提取滑坡位移的空间变形特征，利用时间注意机制和LSTM模型从外部因素的时间序列数据中捕获重要的历史信息.注意力机制输出的注意权重值可以揭示滑坡变形的时间-空间特征.以三峡库区泡桐湾滑坡为例，对该模型的性能进行了验证.结果表明，STA-CNN-LSTM模型预测的均方根误差（RMSE）和平均绝对百分比误差（MAPE）与传统灰狼算法优化的支持向量机（GWO-SVM）模型相比分别下降了9.28%和13.88%.模型因子权重计算结果表明，在监测期内随着时间的推移，降雨对泡桐湾滑坡变形的影响逐渐增加，而库水位的影响逐渐减小.Abstract: Accurate displacement prediction plays an important role in landslide early warning. However, the majority of the existing data-driven models focus on single-point modeling based on time series data which cannot consider the deformation correlation in the whole slope. To overcome this drawback, this study proposed a spatial-temporal attention (STA) mechanism-based deep learning model which combined the convolutional neural network (CNN) with the long short-term memory (LSTM) neural network. A convolutional block attention module (CBAM) combined with CNN was developed to extract the spatial deformation characteristics of the slope. A temporal attention module and LSTM model were used to learn the significant historical information from the input external conditions time series data. The model also allowed to output the tempo-spatial attention weights to reveal the tempo-spatial characteristics of landslide deformation. The Paotongwan landslide with step-like behavior displacement in the Three Gorges Reservoir Area (TGRA) of China was used to validate the model performance. The results show that, the root mean square error (RMSE) and the mean average percentage error (MAPE) of the STA-CNN-LSTM model decreased 9.28% and 13.88%, respectively, compared with grey wolf optimization optimized support vector machine (GWO-SVM). The attention weight results calculated by STA-CNN-LSTM demonstrate that rainfall had a larger impact on the deformation of the Paotongwan landslide at the beginning of the monitoring while the influence of reservoir water level decreased with ongoing of the monitoring.
-
图 2 彩色数字图像（a）和监测网络空间变形数据（b）在CNN中的数据表示
Fig. 2. Presentation of data in CNNs for (a) color digital images and (b) spatial deformation data in monitoring network
图 7 泡桐湾滑坡的累积位移、库水位及降雨量监测数据
灰色矩形表示滑坡变形的4个明显临时加速度
Fig. 7. Monitoring data of cumulative displacement, reservoir level and rainfall at Paotongwan landslide
图 8 泡桐湾滑坡预测位移与实测位移的比较
Fig. 8. The comparison between the predicted and measured displacement at Paotongwan landslide
图 9 泡桐湾滑坡的空间注意权重分布
Fig. 9. Distribution of spatial attention weights of the Paotongwan landslide
图 10 因子注意权重在泡桐湾滑坡上的分布
Fig. 10. Distribution of factor attention weights at the Paotongwan landslide
图 11 泡桐湾滑坡在多个隐藏时间步长下，STA-CNN-LSTM模型与输入因子的可视化时间注意权重比较
Fig. 11. Visualization temporal attention weights compared with input factors of STA-CNN-LSTM model at multiple hidden time steps in Paotongwan landslide
图 12 4种不同模型预测结果与实际值的比较
Fig. 12. Comparison of monitored and predicted landslide displacements using four prediction models for Paotongwan landslide
表 1 不同模型计算的泡桐湾滑坡位移预测结果的误差
Table 1. Errors of the predicted results for the Paotongwan landslide obtained from different models
监测点 RMSE(mm) MAPE(%) GWO-SVM 1CNN-LSTM CNN-LSTM STA-CNN-LSTM GWO-SVM 1CNN-LSTM CNN-LSTM STA-CNN-LSTM WS01 13.16 11.33 10.76 9.06 4.49 4.32 4.10 2.88 WS02 11.57 11.01 12.17 10.13 3.05 3.25 3.81 2.96 WS03 13.52 12.31 13.47 13.56 5.66 4.96 5.48 5.45 WS04 14.18 14.13 11.21 11.18 4.46 5.82 4.75 4.37 WS05 16.87 16.25 16.12 16.17 9.75 7.53 8.15 6.90 WS06 14.09 18.29 13.89 15.02 6.74 9.94 5.83 6.81 平均值 13.90 13.89 13.10 12.61 5.69 5.97 5.35 4.90 
下载: 导出CSV
-
Chen, C. Y., Chen, T. C. C., Yu, W. H., et al., 2005. Rainfall Duration and Debris-Flow Initiated Studies for Real-Time Monitoring. Environmental Geology, 47(5): 715-724. https://doi.org/10.1007/s00254-004-1203-0 Corsini, A., Mulas, M., 2017. Use of ROC Curves for Early Warning of Landslide Displacement Rates in Response to Precipitation (Piagneto Landslide, Northern Apennines, Italy). Landslides, 14(3): 1241-1252. https://doi.org/10.1007/s10346-016-0781-8 Cortes, C., Vapnik, V., 1995. Support-Vector Networks. Machine Learning, 20(3): 273-297. https://doi.org/10.1007/BF00994018 Deng, C. J., Wang, K. W., Yuan, Q. S., et al., 2015. Stability of Reservoir Slope under Repetitive Variation of Reservoir Water Level. Journal of Yangtze River Scientific Research Institute, 32(4): 96-100 (in Chinese with English abstract). doi: 10.3969/j.issn.1001-5485.2015.04.019 Ding, Y. K., Zhu, Y. L., Feng, J., et al., 2020. Interpretable Spatio-Temporal Attention LSTM Model for Flood Forecasting. Neurocomputing, 403: 348-359. https://doi.org/10.1016/j.neucom.2020.04.110 Guo, Z. Z., Chen, L. X., Gui, L., et al., 2020a. Landslide Displacement Prediction Based on Variational Mode Decomposition and WA-GWO-BP Model. Landslides, 17(3): 567-583. https://doi.org/10.1007/s10346-019-01314-4 Guo, Z. Z., Chen, L. X., Yin, K. L., et al., 2020b. Quantitative Risk Assessment of Slow-Moving Landslides from the Viewpoint of Decision-Making: A Case Study of the Three Gorges Reservoir in China. Engineering Geology, 273: 105667. https://doi.org/10.1016/j.enggeo.2020.105667 Hakim, W. L., Rezaie, F., Nur, A. S., et al., 2022. Convolutional Neural Network (CNN) with Metaheuristic Optimization Algorithms for Landslide Susceptibility Mapping in Icheon, South Korea. Journal of Environmental Management, 305: 114367. https://doi.org/10.1016/j.jenvman.2021.114367 Hochreiter, S., Schmidhuber, J., 1997. Long Short-Term Memory. Neural Computation, 9(8): 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735 Huang, D., Gu, D. M., 2017. Influence of Filling-Drawdown Cycles of the Three Gorges Reservoir on Deformation and Failure Behaviors of Anaclinal Rock Slopes in the Wu Gorge. Geomorphology, 295: 489-506. https://doi.org/10.1016/j.geomorph.2017.07.028 Huang, D., Luo, S. L., Zhong, Z., et al., 2020. Analysis and Modeling of the Combined Effects of Hydrological Factors on a Reservoir Bank Slope in the Three Gorges Reservoir Area, China. Engineering Geology, 279: 105858. https://doi.org/10.1016/j.enggeo.2020.105858 Li, S. H., Wu, N., 2021. A New Grey Prediction Model and Its Application in Landslide Displacement Prediction. Chaos, Solitons & Fractals, 147: 110969. https://doi.org/10.1016/j.chaos.2021.110969 Li, X. Z., Kong, J. M., 2014. Application of GA-SVM Method with Parameter Optimization for Landslide Development Prediction. Natural Hazards and Earth System Sciences, 14(3): 525-533. https://doi.org/10.5194/nhess-14-525-2014 Li, Y. J., Xu, Q., He, Y. S., et al., 2017. An ARMA-(LASSO-ELM)-Copula Framework for Landslide Displacement Prediction and Threshold Computing of the Displacement of Step-Like Landslides. Chinese Journal of Rock Mechanics and Engineering, 36(S2): 4075-4084 (in Chinese with English abstract). Liu, Y. Q., Zhang, Q., Song, L. H., et al., 2019. Attention-Based Recurrent Neural Networks for Accurate Short-Term and Long-Term Dissolved Oxygen Prediction. Computers and Electronics in Agriculture, 165: 104964. https://doi.org/10.1016/j.compag.2019.104964 Luo, W. Q., Ji, Y. L., Wang, C. Y., et al., 2016. Research on Seemingly Unrelated Regressions Model of Landslide Displacement Prediction of Multiple Monitoring Points. Chinese Journal of Rock Mechanics and Engineering, 35(S1): 3051-3056 (in Chinese with English abstract). http://www.researchgate.net/publication/303919174_Research_on_seemingly_unrelated_regressions_model_of_landslide_displacement_prediction_of_multiple_monitoring_points Ma, H. T., Zhang, Y. H., Yu, Z. X., 2021. Research on the Identification of Acceleration Starting Point in Inverse Velocity Method and the Prediction of Sliding Time. Chinese Journal of Rock Mechanics and Engineering, 40(2): 355-364 (in Chinese with English abstract). Ma, X. Y., Wang, L. M., Qi, K. L., et al., 2021. Remote Sensing Image Scene Classification Method Based on Multi-Scale Cyclic Attention Network. Earth Science, 46(10): 3740-3752 (in Chinese with English abstract). Ma, Z. J., Mei, G., Prezioso, E., et al., 2021. A Deep Learning Approach Using Graph Convolutional Networks for Slope Deformation Prediction Based on Time-Series Displacement Data. Neural Computing and Applications, 33(21): 14441-14457. https://doi.org/10.1007/s00521-021-06084-6 Miao, F. S., Wu, Y. P., Xie, Y. H., et al., 2017. Research on Progressive Failure Process of Baishuihe Landslide Based on Monte Carlo Model. Stochastic Environmental Research and Risk Assessment, 31(7): 1683-1696. https://doi.org/10.1007/s00477-016-1224-8 Mirjalili, S., Mirjalili, S. M., Lewis, A., 2014. Grey Wolf Optimizer. Advances in Engineering Software, 69: 46-61. https://doi.org/10.1016/j.advengsoft.2013.12.007 Ren, Q. B., Li, M. C., Li, H., et al., 2021. A Novel Deep Learning Prediction Model for Concrete Dam Displacements Using Interpretable Mixed Attention Mechanism. Advanced Engineering Informatics, 50: 101407. https://doi.org/10.1016/j.aei.2021.101407 Shang, M., Liao, F., Ma, R., et al., 2021. Quantitative Correlation Analysis on Deformation of Baijiabao Landslide between Rainfall and Reservoir Water Level. Journal of Engineering Geology, 29(3): 742-750 (in Chinese with English abstract). Song, K., Chen, L. Y., Liu, Y. L., et al., 2022. Dynamic Mechanism of Rain Infiltration in Deep-Seated Landslide Reactivate Deformation. Earth Science, 47(10): 3665-3676 (in Chinese with English abstract). Song, K., Yan, E. C., Zhu, D. P., et al., 2011. Base on Permeability of Landslide and Reservoir Water Change to Research Variational Regularity of Landslide Stability. Rock and Soil Mechanics, 32(9): 2798-2802 (in Chinese with English abstract). doi: 10.3969/j.issn.1000-7598.2011.09.039 Su, X. J., Zhou, W. X., Li, C. J., et al., 2022. Interpretable Offshore Wind Power Output Forecasting Based on Long Short-Term Memory Neural Network with Dual-Stage Attention. Automation of Electric Power Systems, 46(7): 141-151 (in Chinese with English abstract). Wang, Z. W., Li, D. Y., 2005. Application of 3S Technology in Landslide Monitoring. Journal of Yangtze River Scientific Research Institute, 22(5): 33-36 (in Chinese with English abstract). Woo, S., Park, J., 2018. CBAM: Convolutional Block Attention Module. Proceedings of the European Conference on Computer Vision (ECCV), Munich, 3-19. Wu, Q., Zhou, C. B., Huang, F. M., et al., 2022. Optimization of the Landslide Identification Method Based on a Dual Attention Mechanism. Bulletin of Geological Science and Technology, 41(2): 246-253 (in Chinese with English abstract). Xia, M., Ren, G. M., Ma, X. L., 2013. Deformation and Mechanism of Landslide Influenced by the Effects of Reservoir Water and Rainfall, Three Gorges, China. Natural Hazards, 68(2): 467-482. https://doi.org/10.1007/s11069-013-0634-x Xu, C., Liu, B. G., Liu, K. Y., et al., 2011. Intelligent Analysis Model of Landslide Displacement Time Series Based on Coupling PSO-GPR. Rock and Soil Mechanics, 32(6): 1669-1675 (in Chinese with English abstract). doi: 10.3969/j.issn.1000-7598.2011.06.013 Xu, F., Fan, C. J., Xu, X. J., et al., 2018. Displacement Prediction of Landslide Based on Variational Mode Decomposition and AMPSO-SVM Coupling Model. Journal of Shanghai Jiao Tong University, 52(10): 1388-1395, 1416 (in Chinese with English abstract). Xu, S. L., Niu, R. Q., 2018. Displacement Prediction of Baijiabao Landslide Based on Empirical Mode Decomposition and Long Short-Term Memory Neural Network in Three Gorges Area, China. Computers & Geosciences, 111: 87-96. https://doi.org/10.1016/j.cageo.2017.10.013 Yin, Y. P., Huang, B. L., Wang, W. P., et al., 2016. Reservoir-Induced Landslides and Risk Control in Three Gorges Project on Yangtze River, China. Journal of Rock Mechanics and Geotechnical Engineering, 8(5): 577-595. https://doi.org/10.1016/j.jrmge.2016.08.001 Yin, Y. P., Wang, H. D., Gao, Y. L., et al., 2010. Real-Time Monitoring and Early Warning of Landslides at Relocated Wushan Town, the Three Gorges Reservoir, China. Landslides, 7(3): 339-349. https://doi.org/10.1007/s10346-010-0220-1 Zhang, K., Zhang, K., Bao, R., et al., 2021. Intelligent Prediction of Landslide Displacements Based on Optimized Empirical Mode Decomposition and K-Mean Clustering. Rock and Soil Mechanics, 42(1): 211-223 (in Chinese with English abstract). Zheng, J., 2006. Methods and Standards of Landslide Stability Evaluation. The Chinese Journal of Geological Hazard and Control, 17(3): 53-57 (in Chinese with English abstract). Zhou, C., Yin, K. L., Cao, Y., et al., 2018. Displacement Prediction of Step-Like Landslide by Applying a Novel Kernel Extreme Learning Machine Method. Landslides, 15(11): 2211-2225. https://doi.org/10.1007/s10346-018-1022-0 邓成进, 王孔伟, 袁秋霜, 等, 2015. 库水长期升降作用下库岸边坡稳定性研究. 长江科学院院报, 32(4): 96-100. https://www.cnki.com.cn/Article/CJFDTOTAL-CJKB201504022.htm 李骅锦, 许强, 何雨森, 等, 2017. "阶跃式" 滑坡位移预测及阈值分析的ARMA-(LASSO-ELM)-Copula模型. 岩石力学与工程学报, 36(S2): 4075-4084. https://www.cnki.com.cn/Article/CJFDTOTAL-YSLX2017S2041.htm 罗文强, 冀雅楠, 王淳越, 等, 2016. 多监测点滑坡变形预测的似乎不相关模型研究. 岩石力学与工程学报, 35(S1): 3051-3056. https://www.cnki.com.cn/Article/CJFDTOTAL-YSLX2016S1052.htm 马海涛, 张亦海, 于正兴, 2021. 滑坡速度倒数法预测模型加速开始点识别及临滑时间预测研究. 岩石力学与工程学报, 40(2): 355-364. https://www.cnki.com.cn/Article/CJFDTOTAL-YSLX202102011.htm 马欣悦, 王梨名, 祁昆仑, 等, 2021. 基于多尺度循环注意力网络的遥感影像场景分类方法. 地球科学, 46(10): 3740-3752. doi: 10.3799/dqkx.2020.365 尚敏, 廖芬, 马锐, 等, 2021. 白家包滑坡变形与库水位、降雨相关性定量化分析研究. 工程地质学报, 29(3): 742-750. https://www.cnki.com.cn/Article/CJFDTOTAL-GCDZ202103017.htm 宋琨, 陈伦怡, 刘艺梁, 等, 2022. 降雨诱发深层老滑坡复活变形的动态作用机制. 地球科学, 47(10): 3665-3676. doi: 10.3799/dqkx.2022.184 宋琨, 晏鄂川, 朱大鹏, 等, 2011. 基于滑体渗透性与库水变动的滑坡稳定性变化规律研究. 岩土力学, 32(9): 2798-2802. https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX201109043.htm 苏向敬, 周汶鑫, 李超杰, 等, 2022. 基于双重注意力LSTM神经网络的可解释海上风电出力预测. 电力系统自动化, 46(7): 141-151. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT202207016.htm 王志旺, 李端有, 2005.3S技术在滑坡监测中的应用. 长江科学院院报, 22(5): 33-36. https://www.cnki.com.cn/Article/CJFDTOTAL-CJKB200505010.htm 吴琪, 周创兵, 黄发明, 等, 2022. 基于双重注意力机制的滑坡识别方法优化. 地质科技通报, 41(2): 246-253. https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202202025.htm 徐冲, 刘保国, 刘开云, 等, 2011. 基于粒子群-高斯过程回归耦合算法的滑坡位移时序分析预测智能模型. 岩土力学, 32(6): 1669-1675. 徐峰, 范春菊, 徐勋建, 等, 2018. 基于变分模态分解和AMPSO-SVM耦合模型的滑坡位移预测. 上海交通大学学报, 52(10): 1388-1395, 1416. https://www.cnki.com.cn/Article/CJFDTOTAL-SHJT201810031.htm 张凯, 张科, 保瑞, 等, 2021. 基于优化经验模态分解和聚类分析的滑坡位移智能预测研究. 岩土力学, 42(1): 211-223. https://www.cnki.com.cn/Article/CJFDTOTAL-YTLX202101024.htm 郑静, 2006. 滑坡稳定性评价的方法及标准. 中国地质灾害与防治学报, 17(3): 53-57. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGDH200603014.htm -
点击查看大图
图(12) / 表(1)
计量
- 文章访问数: 668
- HTML全文浏览量: 601
- PDF下载量: 77
- 被引次数: 0
