A spatiotemporal deep learning method for excavation-induced wall deflections

Data-driven approaches such as neural networks are increasingly used for deep excavations due to the growing amount of available monitoring data in practical projects.However,most neural network models only use the data from a single monitoring point and neglect the spatial relationships between multiple monitoring points.Besides,most models lack flexibility in providing predictions for multiple days after monitoring activity.This study proposes a sequence-to-sequence(seq2seq)two-dimensional(2D)convolutional long short-term memory neural network(S2SCL2D)for predicting the spatiotemporal wall deflections induced by deep excavations.The model utilizes the data from all monitoring points on the entire wall and extracts spatiotemporal features from data by combining the 2D convolutional layers and long short-term memory(LSTM)layers.The S2SCL2D model achieves a long-term prediction of wall deflections through a recursive seq2seq structure.The excavation depth,which has a significant impact on wall deflections,is also considered using a feature fusion method.An excavation project in Hangzhou,China,is used to illustrate the proposed model.The results demonstrate that the S2SCL2D model has superior prediction accuracy and robustness than that of the LSTM and S2SCL1D(one-dimensional)models.The prediction model demonstrates a strong generalizability when applied to an adjacent excavation.Based on the long-term prediction results,practitioners can plan and allocate resources in advance to address the potential engineering issues.

Deterministic and probabilistic analysis of great-depth braced excavations:A 32 m excavation case study in Paris

The Fort d’Issy-Vanves-Clamart(FIVC)braced excavation in France is analyzed to provide insights into the geotechnical serviceability assessment of excavations at great depth within deterministic and probabilistic frameworks.The FIVC excavation is excavated at 32 m below the ground surface in Parisian sedimentary basin and a plane-strain finite element analysis is implemented to examine the wall deflections and ground surface settlements.A stochastic finite element method based on the polynomial chaos Kriging metamodel(MSFEM)is then proposed for the probabilistic analyses.Comparisons with field measurements and former studies are carried out.Several academic cases are then conducted to investigate the great-depth excavation stability regarding the maximum horizontal wall deflection and maximum ground surface settlement.The results indicate that the proposed MSFEM is effective for probabilistic analyses and can provide useful insights for the excavation design and construction.A sensitivity analysis for seven considered random parameters is then implemented.The soil friction angle at the excavation bottom layer is the most significant one for design.The soil-wall interaction effects on the excavation stability are also given.

Influence of groundwater drawdown on excavation responses e A case history in Bukit Timah granitic residual soils

Performances of a braced cut-and-cover excavation system for mass rapid transit （MRT） stations of the Downtown Line Stage 2 in Singapore are presented. The excavation was carried out in the Bukit Timah granitic （BTG） residual soils and characterized by significant groundwater drawdown, due to dewatering work in complex site conditions, insufficient effective waterproof measures and more permeable soils. A two-dimensional numerical model was developed for back analysis of retaining wall movement and ground surface settlement. Comparisons of these measured excavation responses with the calculated performances were carried out, upon which the numerical simulation procedures were calibrated. In addition, the influences of groundwater drawdown on the wall deflection and ground surface settlement were numerically investigated and summarized. The performances were also compared with some commonly used empirical charts, and the results indicated that these charts are less applicable for cases with significant groundwater drawdowns. It is expected that these general behaviors will provide useful references and insights for future projects involving excavation in BTG residual soils under significant groundwater drawdowns.

An intelligent procedure for updating deformation prediction of braced excavation in clay using gated recurrent unit neural networks

This paper aims to establish an intelligent procedure that combines the observational method with the existing deep learning technique for updating deformation of braced excavation in clay.The gated recurrent unit(GRU) neural network is adopted to formulate the forecast model and learn the potential rules in the field observations using the Nesterov-accelerated Adam(Nadam) algorithm.In the proposed procedure,the GRU-based forecast model is first trained based on the field data of previous and current stages.Then,the field data of the current stage are used as input to predict the deformation response of the next stage via the previously trained GRU-based forecast model.This updating process will loop up till the end of the excavation.This procedure has the advantage of directly predicting the deformation response of unexcavated stages based on the monitoring data.The proposed intelligent procedure is verified on two well-documented cases in terms of accuracy and reliability.The results indicate that both wall deflection and ground settlement are accurately predicted as the excavation proceeds.Furthermore,the advantages of the proposed intelligent procedure compared with the Bayesian/o ptimization updating are illustrated.

Optimized functional linked neural network for predicting diaphragm wall deflection induced by braced excavations in clays

Deep excavation during the construction of underground systems can cause movement on the ground,especially in soft clay layers.At high levels,excessive ground movements can lead to severe damage to adjacent structures.In this study,finite element analyses(FEM)and the hardening small strain(HSS)model were performed to investigate the deflection of the diaphragm wall in the soft clay layer induced by braced excavations.Different geometric and mechanical properties of the wall were investigated to study the deflection behavior of the wall in soft clays.Accordingly,1090 hypothetical cases were surveyed and simulated based on the HSS model and FEM to evaluate the wall deflection behavior.The results were then used to develop an intelligent model for predicting wall deflection using the functional linked neural network(FLNN)with different functional expansions and activation functions.Although the FLNN is a novel approach to predict wall deflection;however,in order to improve the accuracy of the FLNN model in predicting wall deflection,three swarm-based optimization algorithms,such as artificial bee colony(ABC),Harris’s hawk’s optimization(HHO),and hunger games search(HGS),were hybridized to the FLNN model to generate three novel intelligent models,namely ABC-FLNN,HHO-FLNN,HGS-FLNN.The results of the hybrid models were then compared with the basic FLNN and MLP models.They revealed that FLNN is a good solution for predicting wall deflection,and the application of different functional expansions and activation functions has a significant effect on the outcome predictions of the wall deflection.It is remarkably interesting that the performance of the FLNN model was better than the MLP model with a mean absolute error(MAE)of 19.971,root-mean-squared error(RMSE)of 24.574,and determination coefficient(R^(2))of 0.878.Meanwhile,the performance of the MLP model only obtained an MAE of 20.321,RMSE of 27.091,and R^(2)of 0.851.Furthermore,the results also indicated that the proposed hybrid models,i.e.,ABC-FLNN,HHO-FLNN,HGS-FLNN,yielded more superior performances than those of the FLNN and MLP models in terms of the prediction of deflection behavior of diaphragm walls with an MAE in the range of 11.877 to 12.239,RMSE in the range of 15.821 to 16.045,and R^(2)in the range of 0.949 to 0.951.They can be used as an alternative tool to simulate diaphragm wall deflections under different conditions with a high degree of accuracy.

Behavior of braced excavation in sand under a seismic condition: experimental and numerical studies

The behavior of braced excavation in dry sand under a seismic condition is investigated in this paper.A series of shake table tests on a reduced scale model of a retaining wall with one level of bracing were conducted to study the effect of different design parameters such as excavation depth,acceleration amplitude and wall stiffness.Numerical analyses using FLAC 2D were also performed considering one level of bracing.The strut forces,lateral displacements and bending moments in the wall at the end of earthquake motion were compared with experimental results.The study showed that in a post-seismic condition,when other factors were constant,lateral displacement,bending moment,strut forces and maximum ground surface displacement increased with excavation depth and the amplitude of base acceleration.The study also showed that as wall stiffness decreased,the lateral displacement of the wall and ground surface displacement increased,but the bending moment of the wall and strut forces decreased.The net earth pressure behind the walls was influenced by excavation depth and the peak acceleration amplitude,but did not change significantly with wall stiffness.Strut force was the least affected parameter when compared with others under a seismic condition.

Base stability analysis of braced deep excavation in undrained anisotropic clay with upper bound theory

It is imperative to evaluate factor of safety against basal heave failure in the design of braced deep excavation in soft clay.Based on previously published field monitoring data and finite element analyses of ground settlements of deep excavation in soft clay,an assumed plastic deformation mechanism proposed here gives upper bound solutions for base stability of braced deep excavations.The proposed kinematic mechanism is optimized by the mobile depth(profile wavelength).The method takes into account the influence of strength anisotropy under plane strain conditions,the embedment of the retaining wall,and the locations of the struts.The current method is validated by comparison with published numerical study of braced excavations in Boston blue clay and two other cases of excavation failure in Taipei.The results show that the upper bound solutions obtained from this presented method is more accurate as compared with the conventional methods for basal heave failure analyses.

Performance of braced excavation in residual soil with groundwater drawdown

In densely built-up Singapore,relatively stiffsecant-bored piles and diaphragm walls are commonly used in cut-and-cover works to minimize the impact of ground movement on the adjacent structures and utilities.For excavations in stiffresidual soil deposits,the asso-ciated wall deflections and ground settlements are generally smaller than for excavations in soft soil deposits.However,if the residual soil permeability is high and the underlying rock is highlyfissured or fractured,substantial groundwater drawdown and associated seepage-induced settlement may occur.In this study,the excavation performance of four sites in residual soil deposits with maximum excavation depths between 20 and 24 m is presented.The maximum wall deflections were found to be relatively small compared to the significantly larger maximum ground settlements,owing to the extensive lowering of the groundwater table.In this paper,details of the subsurface conditions,excavation support system,field instrumentation,and observed excavation responses are presented,with particular focus on the large groundwater drawdown and associated ground settlement.Specific issues encountered during the excavation,as well as the effectiveness of various groundwater control measures,are discussed.The case studies will provide useful references and insights for future projects involving braced excavations in residual soil.

Effects of soil spatial variability on structural reliability assessment in excavations

In this study,the effects of soil spatial variability in braced excavations are investigated by focusing on three structural responses:wall bending moments,wall shear forces,and strut forces.The soil spatial variability is modeled using random field theory,and the generated soil parameters are mapped onto a finite element model.A procedure for automating the Monte Carlo simulation,which is used for probabilistic analysis,is described.A case study demonstrates that the soil spatial variability has a considerable effect on the excavation-induced structural responses.Furthermore,a reliability analysis is performed to estimate the failure probability for three structural failure modes.The results demonstrate the importance of considering soil spatial variability in the structural assessment of braced excavati ons.

Estimation of failure probability in braced excavation using Bayesian networks with integrated model updating

A probabilistic model is proposed that uses observation data to estimate failure probabilities during excavations.The model integrates a Bayesian network and distanced-based Bayesian model updating.In the network,the movement of a retaining wall is selected as the indicator of failure,and the observed ground surface settlement is used to update the soil parameters.The responses of wall deflection and ground surface settlement are accurately predicted using finite element analysis.An artificial neural network is employed to construct the response surface relationship using the aforementioned input factors.The proposed model effectively estimates the uncertainty of influential factors.A case study of a braced excavation is presented to demonstrate the feasibility of the proposed approach.The update results facilitate accurate estimates according to the target value,from which the corresponding probabilities of failure are obtained.The proposed model enables failure probabilities to be determined with real-time result updating.

Assessment of basal heave stability for braced excavations in anisotropic clay using extreme gradient boosting and random forest regression

A finite-element analysis considering the anisotropy for the undrained shear strength was performed to examine the effects of the total stress-based anisotropic model NGI-ADP(developed by Norwegian Geotechnical Institute based on the Active-Direct simple shear-Passive concept)parameters on the base stability of deep braced excavations in clays.These parameters included the ratio of the plane strain passive shear strength to the plane strain active shear strength s_(u)^(P)=s_(u)^(A),the ratio of the unloading/reloading shear modulus to the plane strain active shear strength G_(ur)=s_(u)^(A),the plane strain active shear strength s_(u)^(A),the unit weight c,the excavation width B,the wall thickness b,and the wall penetration depth D.According to the numerical results for 1778 hypothetical cases,extreme gradient boosting(XGBoost)and random forest regression(RFR)were adopted to predict the factor of safety(FS)against basal heave for deep braced excavations.The results indicated that the anisotropic characteristics of soil parameters need to be considered when determining the FS against basal heave for braced excavation.XGBoost and RFR can yield a reasonable prediction of the FS.This paper presents a cuttingedge application of ensemble learning methods in geotechnical engineering.