测绘学报 ›› 2024, Vol. 53 ›› Issue (6): 1128-1139.doi: 10.11947/j.AGCS.2024.20230463

• 智能化测绘 • 上一篇    下一篇

机理引导下的阶跃型滑坡位移预测深度学习模型

蒋亚楠1,2,3(), 郑林枫1, 许强2(), 汤明高2, 朱星2,3   

  1. 1.成都理工大学地球与行星科学学院,四川 成都 610059
    2.成都理工大学地质灾害防治与地质环境保护国家重点实验室,四川 成都 610059
    3.四川省工业互联网智能监测及应用工程技术研究中心,四川 成都 610059
  • 收稿日期:2023-10-07 发布日期:2024-07-22
  • 通讯作者: 许强 E-mail:jiangyanan@cdut.edu.cn;xq@cdut.edu.cn
  • 作者简介:蒋亚楠(1988—),女,博士,副教授,研究方向为地质灾害早期识别与监测预警。 E-mail:jiangyanan@cdut.edu.cn
  • 基金资助:
    长江生态环境保护修复联合研究二期项目(2022-LHYJ-02-0201);国家自然科学基金(42304042);四川省重点研发项目(2023YFS0439)

Step-like displacement prediction of landslides guided by deformation mechanism

Yanan JIANG1,2,3(), Linfeng ZHENG1, Qiang XU2(), Minggao TANG2, Xing ZHU2,3   

  1. 1.School of Earth and Planetary Sciences, Chengdu University of Technology, Chengdu 610059, China
    2.State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, Chengdu 610059, China
    3.Sichuan Engineering Technology Research Center of Industrial Internet Intelligent Monitoring and Application, Chengdu 610059, China
  • Received:2023-10-07 Published:2024-07-22
  • Contact: Qiang XU E-mail:jiangyanan@cdut.edu.cn;xq@cdut.edu.cn
  • About author:JIANG Yanan (1988—), female, PhD, associate professor, majors in early identification, monitoring and warning of geological hazards. E-mail: jiangyanan@cdut.edu.cn
  • Supported by:
    The Yangtze River Joint Research Phase Ⅱ Program(2022-LHYJ-02-0201);The National Science Foundation of China(42304042);The Key Research and Development Program of Sichuan Province(2023YFS0439)

摘要:

阶跃型滑坡变形时间曲线呈阶梯状,阶跃变形量大,准确预警预报困难。针对现有模型在阶跃型滑坡快速变形阶段预测误差大的问题,提出一种机理引导下的阶跃型滑坡变形预测模型,该模型在深入分析滑坡变形机理上,结合变分模态分解开展滑坡位移和影响因子的动态响应分析,为Informer模型提供合理有效的外部影响因子输入,结合多头注意力机制和池化层,实现阶跃期时序数据关键周期信息的有效提取。本研究以三峡库区白水河滑坡为例,收集水库蓄水以来连续15年的逐月位移监测数据及同期逐天的降雨和库水位数据。试验结果表明,本文模型在阶跃型滑坡位移预测中整体预测精度较高,与主流预测模型相比,该模型对快速变形期的阶跃变形预测较为准确,预测误差较小。

关键词: 阶跃型滑坡, 变形机理, 位移预测, Informer, 自注意力机制, 影响因子

Abstract:

Rainfall reservoir-induced landslides in the Three Gorges Reservoir (TGR), China, exhibit distinctive step-like deformation characteristics, involving mutation and creep states. These particular features pose a challenge for accurate early warning and prediction. Previous landslide displacement forecasting models have shown limited prediction accuracy, particularly when it comes to mutational displacements. The proposed prediction model in this study, based on Informer, utilizes a multi-head attention mechanism to capture temporal dependencies and incorporates pooling layers for emphasizing crucial features, enabling adaptive learning of feature weights and more effective extraction of periodic information from time series data. The Baishuihe landslide was used for case studies with monitoring data collected from July 2013 to December 2018, including monthly displacements, daily rainfall and reservoir water level. Firstly, cumulative displacement was decomposed into trend displacement and periodic displacement by the variational mode decomposition (VMD). After triggering factors selection and decomposition, the double exponential smoothing (DES) method and the Informer model are used to predict the trend and periodic component displacements, respectively. Finally, the predicted trend and periodic components are combined to generate the cumulative displacement prediction. Results demonstrate that the proposed model achieves impressive results with a root mean square error of 12.21 mm, a mean absolute error of 10.05 mm, and a coefficient of determination of 0.99 for the next 27 months' cumulative displacement prediction. Compared to other four mainstream models, this approach exhibits higher prediction accuracy, particularly in predicting the rapid deformation phase of step-like bank landslides. Consequently, it holds significant credibility and practical value in the early warning research of rainfall reservoir-induced landslides.

Key words: step-like landslide, deformation mechanism, displacement prediction, Informer, attention mechanism, triggering factors

中图分类号: