Displacement prediction of water-induced landslides using a recurrent deep learning model

Meng, Qingxiang; Wang, Huanling; He, Mingjie; Gu, Jing; Qi, Ji; Yang, Liang

doi:10.1080/19648189.2020.1763847

Cited by 22 publications

(14 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the ML literature, the RNN models such as the long shortterm memory (LSTM) have been developed for forecasting soil movements (Xing et al, 2019;Yang et al, 2019;Jiang et al, 2020;Liu et al, 2020;Meng et al, 2020;Niu et al, 2021). These recurrent models possess internal memory and they are a generalization of the feedforward neural networks (Medsker and Jain, 1999).…”

Section: Introductionmentioning

confidence: 99%

BS-LSTM: An Ensemble Recurrent Approach to Forecasting Soil Movements in the Real World

et al. 2021

View full text Add to dashboard Cite

Machine learning (ML) proposes an extensive range of techniques, which could be applied to forecasting soil movements using historical soil movements and other variables. For example, researchers have proposed recurrent ML techniques like the long short-term memory (LSTM) models for forecasting time series variables. However, the application of novel LSTM models for forecasting time series involving soil movements is yet to be fully explored. The primary objective of this research is to develop and test a new ensemble LSTM technique (called “Bidirectional-Stacked-LSTM” or “BS-LSTM”). In the BS-LSTM model, forecasts of soil movements are derived from a bidirectional LSTM for a period. These forecasts are then fed into a stacked LSTM to derive the next period’s forecast. For developing the BS-LSTM model, datasets from two real-world landslide sites in India were used: Tangni (Chamoli district) and Kumarhatti (Solan district). The initial 80% of soil movements in both datasets were used for model training and the last 20% of soil movements in both datasets were used for model testing. The BS-LSTM model’s performance was compared to other LSTM variants, including a simple LSTM, a bidirectional LSTM, a stacked LSTM, a CNN-LSTM, and a Conv-LSTM, on both datasets. Results showed that the BS-LSTM model outperformed all other LSTM model variants during training and test in both the Tangni and Kumarhatti datasets. This research highlights the utility of developing recurrent ensemble models for forecasting soil movements ahead of time.

show abstract

Section: Introductionmentioning

confidence: 99%

BS-LSTM: An Ensemble Recurrent Approach to Forecasting Soil Movements in the Real World

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Considering the basic properties of miscellaneous fill, scholars in China and foreign countries have discussed its mechanical properties through laboratory experiments. Some experimental instruments, such as large-and middlesized triaxial apparatus [4][5][6][7][8][9][10], shear apparatus [11,12], and horizontal push-shear text [13][14][15][16], are commonly used. Existing associated studies focus on moisture content and strength characteristics of miscellaneous fill [17,18], compression [19], permeability, and particle composition.…”

Section: Introductionmentioning

confidence: 99%

Numerical Simulation of Mutually Embedded Settlement in Miscellaneous Fill

Zhang

Liu

Chen

et al. 2020

Advances in Civil Engineering

View full text Add to dashboard Cite

Embedding soft soil particles with high moisture content into miscellaneous fill with large pores under overlying loads is easy. It produces mutually embedded settlement, which is an important component of total foundation settlement during calculation. In this study, influences of interface friction on mutually embedded settlement, particle displacement, pore and contact characteristics, and mutually embedded development laws were analysed by using the particle flow method. Research results demonstrate that mutually embedded settlement decreases first and then stabilizes with the increase in contact friction factor and continuously attenuates with normal stiffness. Under the loads, particles at the contact surface move downward and squeeze surrounding particles laterally, thus causing particles to slide at the miscellaneous fill channel upward. Consequently, porosity of particles in miscellaneous fill channel increases. The force chain at the contact surface inclines around, while that at the miscellaneous fill channel presents approximately horizontal distribution. Compared with 35 and 45 mm particles, the mutually embedded settlement of 15 and 25 mm particles is slightly increased with loads. Particle size can relieve the influences of loads on mutual embedding. When particle size is larger than 25 mm, loads can significantly influence mutual embedding. Research conclusions can provide a reasonable theoretical foundation for calculating or predicting settlement of miscellaneous fill-soft soil composite foundation.

show abstract

“…However, many traditional SWRC models [6][7][8][9][10][11] do not consider the effect of porosity on SWRC, which may cause an inaccurate result. Therefore, it needs to establish the relationship between porosity and SWRC in agricultural engineering and geotechnical engineering, especially involving water-induced landslides [12][13][14] and water-rock coupling engineering [15][16][17].…”

Section: Introductionmentioning

confidence: 99%

Effect of Porosity on Soil-Water Retention Curves: Theoretical and Experimental Aspects

et al. 2020

View full text Add to dashboard Cite

Porosity change is a common characteristic of natural soils in fluid-solid interaction problems, which can lead to an obvious change of the soil-water retention curve (SWRC). The influence of porosity on soil water retention phenomena is investigated by a theoretical model and an experimental test in this study. A model expressing the change in suction with porosity and effective saturation is put forward theoretically. The model is based on an idealization of three-phase porous materials, the pore structures of which are homogeneous and isotropic. It accounts for the porosity effect on soil water retention, using four parameters with clear physical meanings. The presented model can obtain the SWRC at any porosity, which will reduce the test number required in characterizing the hydraulic behavior of soil. A laboratory experiment for loamy sand with different porosities is performed. The test results show that suction has a significant variation with changes in porosity and decreases with the increase of porosity. The formulation is verified by both the test data and the literature data for FEBEX bentonite and Boom clay. The very good agreements between measured and predicted results show that the SWRC model is reliable and feasible for various soils.

show abstract

Displacement prediction of water-induced landslides using a recurrent deep learning model

Cited by 22 publications

References 33 publications

BS-LSTM: An Ensemble Recurrent Approach to Forecasting Soil Movements in the Real World

BS-LSTM: An Ensemble Recurrent Approach to Forecasting Soil Movements in the Real World

Numerical Simulation of Mutually Embedded Settlement in Miscellaneous Fill

Effect of Porosity on Soil-Water Retention Curves: Theoretical and Experimental Aspects

Contact Info

Product

Resources

About