An efficient physics-guided Bayesian framework for predicting ground settlement profile during excavations in clay

Recently,the application of Bayesian updating to predict excavation-induced deformation has proven successful and improved prediction accuracy significantly.However,updating the ground settlement profile,which is crucial for determining potential damage to nearby infrastructures,has received limited attention.To address this,this paper proposes a physics-guided simplified model combined with a Bayesian updating framework to accurately predict the ground settlement profile.The advantage of this model is that it eliminates the need for complex finite element modeling and makes the updating framework user-friendly.Furthermore,the model is physically interpretable,which can provide valuable references for construction adjustments.The effectiveness of the proposed method is demonstrated through two field case studies,showing that it can yield satisfactory predictions for the settlement profile.

Ground settlement and tunnel response due to twin-curved shield tunnelling in soft ground with small clear distance

Twin curved tunnels are often encountered in shield tunnelling,where significant complexities in densely exploited underground space are observed.In this study,the ground settlement and tunnel deformation due to twin-curved shield tunnelling in soft ground were investigated using numerical simulation and field monitoring.Different curvature radii of twin curved tunnels and subsequent effects of tunnel construction were considered to reveal the tunnelling effect on ground surface settlement and tunnel deformation.The results show that the settlement trough yields one offset towards inside of curved shield tunnelling.The location of settlement trough and maximum settlement were affected by curvature radius but except for the shape and width of settlement trough.Adjacent parallel twin-curved shield tunnelling could increase the offset of existing settlement trough and maximum settlement.Then,an empirical prediction of surface settlement trough due to twin-curved shield tunnelling with same tunnel diameters in soft clay was proposed,which was applicable to curvature radius less than 800 m.Finally,a minimum radius of 600 m of curvature tunnel was proposed in terms of allowable convergence deformation of tunnel.The result could provide guidance on safety evaluation for twin curved shield tunnelling construction.

Effectiveness of predicting tunneling-induced ground settlements using machine learning methods with small datasets

Prediction of tunneling-induced ground settlements is an essential task,particularly for tunneling in urban settings.Ground settlements should be limited within a tolerable threshold to avoid damages to aboveground structures.Machine learning(ML)methods are becoming popular in many fields,including tunneling and underground excavations,as a powerful learning and predicting technique.However,the available datasets collected from a tunneling project are usually small from the perspective of applying ML methods.Can ML algorithms effectively predict tunneling-induced ground settlements when the available datasets are small?In this study,seven ML methods are utilized to predict tunneling-induced ground settlement using 14 contributing factors measured before or during tunnel excavation.These methods include multiple linear regression(MLR),decision tree(DT),random forest(RF),gradient boosting(GB),support vector regression(SVR),back-propagation neural network(BPNN),and permutation importancebased BPNN(PI-BPNN)models.All methods except BPNN and PI-BPNN are shallow-structure ML methods.The effectiveness of these seven ML approaches on small datasets is evaluated using model accuracy and stability.The model accuracy is measured by the coefficient of determination(R2)of training and testing datasets,and the stability of a learning algorithm indicates robust predictive performance.Also,the quantile error(QE)criterion is introduced to assess model predictive performance considering underpredictions and overpredictions.Our study reveals that the RF algorithm outperforms all the other models with the highest model prediction accuracy(0.9)and stability(3.0210^(-27)).Deep-structure ML models do not perform well for small datasets with relatively low model accuracy(0.59)and stability(5.76).The PI-BPNN architecture is proposed and designed for small datasets,showing better performance than typical BPNN.Six important contributing factors of ground settlements are identified,including tunnel depth,the distance between tunnel face and surface monitoring points(DTM),weighted average soil compressibility modulus(ACM),grouting pressure,penetrating rate and thrust force.

Failure mechanism of large-diameter shield tunnels and its effects on ground surface settlements

A new technique for the analysis of the three-dimensional collapse failure mechanism and the ground surface settlements for the large-diameter shield tunnels were presented.The technique is based on a velocity field model using more different truncated solid conical blocks to clarify the multiblock failure mechanism.Furthermore,the shape of blocks between the failure surface and the tunnel face was considered as an entire circle,and the supporting pressure was assumed as non-uniform distribution on the tunnel face and increased with the tunnel embedded depth.The ground surface settlements and failure mechanism above large-diameter shield tunnels were also investigated under different supporting pressures by the finite difference method.

Analysis of ground surface settlement in anisotropic clays using extreme gradient boosting and random forest regression models

Excessive ground surface settlement induced by pit excavation(i.e.braced excavation) can potentially result in damage to the nearby buildings and facilities.In this paper,extensive finite element analyses have been carried out to evaluate the effects of various structural,soil and geometric properties on the maximum ground surface settlement induced by braced excavation in anisotropic clays.The anisotropic soil properties considered include the plane strain shear strength ratio(i.e.the ratio of the passive undrained shear strength to the active one) and the unloading shear modulus ratio.Other parameters considered include the support system stiffness,the excavation width to excavation depth ratio,and the wall penetration depth to excavation depth ratio.Subsequently,the maximum ground surface settlement of a total of 1479 hypothetical cases were analyzed by various machine learning algorithms including the ensemble learning methods(extreme gradient boosting(XGBoost) and random forest regression(RFR)algorithms).The prediction models developed by the XGBoost and RFR are compared with that of two conventional regression methods,and the predictive accuracy of these models are assessed.This study aims to highlight the technical feasibility and applicability of advanced ensemble learning methods in geotechnical engineering practice.

Analysis of ground deformation development and settlement prediction by air-boosted vacuum preloading

Vacuum preloading has been widely used to improve soft soils in coastal areas of China.An increasing amount of evidence from field operations has shown that conventional vacuum preloading is prone to clogging in prefabricated vertical drains(PVDs)and demands a large volume of sand fills.In recent years,air-boosted vacuum preloading has been developed to overcome these limitations;however,this method still requires more data to verify its performance.In this study,a field test for air-boosted vacuum preloading was conducted,and a large-strain two-dimensional(2D)finite element(FE)model was developed and validated against the field test data.Then,a series of FE parametric analyses was performed to assess key factors,i.e.the air injection pressure,the injection spacing,and the characteristics of cyclic injection,which affect the performance of the air-boosted vacuum preloading.The results showed that the ground settlement and lateral displacement of the soils increased due to an increase in the injection pressure,a decrease in the injection spacing,or increases in the number and duration of the injection cycles.Based on the parametric analyses,an empirical formula for ground settlement prediction was proposed and compared with a case history reported in the literature,showing good agreement.

Study on influence of ground settlement in bored metro tunnel

At present the bored construction method is one of the main construction methods of metro and tunnel construction in China. The empirical estimated formulas of tunnel ground surface settlement using the bored construction method were obtained, combining the mechanical stimulant calculated result of tunnel model of different embedded depth, different cross section and different construction method and the actual measurement data of Beijing metro construction. According to the regressed analysis of calculated data, the calculated equations of ground surface settlement value and settlement range of tunnel section under the condition of different embedded depth, different cross section and different construction method were gained. Among them there are some empirical formulas can apply to the construction design of metro tunnel directly.

Measured Performance and Analysis of the Residual Settlement of a PVD-Improved Marine Soft Ground

Prefabricated vertical drains(PVDs)are commonly used to shorten the drainage path for consolidation as part of the improvement of marine soft ground.Many studies that focus on the primary consolidation settlement of PVD-improved soft ground have been conducted;however,residual settlement has been scarcely investigated.Residual settlement is the net effect of secondary compression and the remaining primary consolidation and generally occurs while the facilities are operating.In this study,residual settlement was investigated using the measured field settlement data obtained from the surface settlement plate and multilayer settlement gauges.This study determined that PVD still has some effect on residual settlement and can reduce the settlement times.Residual settlement is only related to the PVD-improved soil layer and only occurs significantly in the middle zone of that layer over a few months.The middle zone may be related to the time delay of excess pore water pressure dissipation.This study concluded that the remaining primary consolidation in the PVD-improved soil layer is the primary cause of residual settlement,whereas secondary compression in the PVD-improved soil layer is only a minor cause.

Ground settlement sedimentation time-space effect and ground settlement control standards

Since different underground engineering contains different geological conditions,structure and excavation methods,their disturbing degree on the ground also differs,but ground surface settlement caused by excavation has one common point that they will form a ground surface settlement curve.Based on data statistical analysis,the article puts forward the relationship between point of inflection in digging tunnels with shallowburied method and the span of tunnel excavation against typical geological conditions in Changchun so as to predict the impact of tunnel excavation and improve reasonable ground surface settlement control standard.Research result will be useful for study on ground surface control standard in digging tunnels with shallow-buried method and setting settlement standard.

Shallow Tunneling Method and Control Measure for Ground Surface Settlement

In order to ensure that the tunnel deformation and surface settlement are controlled within the allowable range during the construction process,the design unit has compiled technical measures and monitoring schemes for ground settlement control of this project.Based on the example of a shallow tunneling project on Subway line 8,this paper analyzes and discusses the shallow tunneling method in detail and puts forward corresponding technical measures for ground settlement control.

Modified image analytical solutions for ground displacement using nonuniform convergence model

Based on the image theory,the analytical solutions of tunneling-induced ground displacement were derived in conjunction with the nonuniform convergence model.The reasonable value of Poisson ratio in the analytical solution was discussed.The ground settlement width parameter which could reflect the ground condition was introduced to modify the analytical solutions proposed above,and new analytical solutions were presented.To evaluate the validity of the present solutions using the nonuniform convergence model,the results were compared with the observed values for four engineering projects,including 38 measured data of ground settlement.The agreement shows that the present solutions using the nonuniform convergence model are effective for evaluating the tunneling-induced ground displacements.

Relationships between ground motion parameters and landslides induced by Wenchuan earthquake

The MS8.0 Wenchuan earthquake induced severe landslide hazards.For the first time in China,large numbers of strong motion records were obtained during the Wenchuan earthquake,providing the opportunity to study the relationships between ground-motion parameters and the earthquake-induced landslides.Nearly 40 groups of records from the main shock distributed along the Longmenshan fault lines were used to carry out this study.The results appropriate to the Longmenshan area are as follows：1 The threshold of the peak ground acceleration（PGA） is about 0.7 m/s2.When the PGA reaches 2 m/s2,the landslide hazards are very serious; 2 The threshold of the peak ground velocity（PGV） is about 0.5 m/s.When the PGV reaches 1.5 m/s,severe landslide hazards will be induced; 3 The threshold for the Arias intensity（Ia） is about 0.2 m/s.When the Ia in one horizontal direction reaches 2 m/s,landslide hazards will be very serious; 4 As for the relevance order of the parameters to earthquake-induced landslides,Ia is the leading parameter,followed by PGV,and finally PGA.The results presented in this paper are consistent with the results from other studies,indicating that the threshold of the ground motion parameters for strong earthquakes is of the same order of magnitude as that of moderate earthquakes.Landslide density of local sites fluctuated with the increase of ground motion intensity if the thresholds were reached.When the upper limits are exceeded,the landslide density remains at a certain level with relatively little variation.

Centrifuge model test on earthquake-induced differential settlement of foundation on cohesive ground

Dynamic centrifuge model test was conducted to study the earthquake-induced differential settlement of foundation on cohesive ground, and the influence of asymmetry of building was investigated. During the experiment, the overconsolidated kaolin clay ground with a three-dimensional asymmetrical structure model was shaken by a basically balanced input motion, and bender elements were used to measure shear wave velocities of model ground to reveal the soil fabric evolution during and after shaking. The test results show that, the total seismic settlement of foundation is composed of instantaneous and long-term post-earthquake settlements, and most of the differential settlement occurs immediately after the earthquake while the post-earthquake settlement is relatively uniform despite its large amplitude. The asymmetry of building affects the settlement behavior considerably. Compared with 1-or 2-dimensional structures, more evident differential settlement occurs under threedimensional asymmetrical building during shaking, which accounts for one-half of the total seismic settlements and results in complex spatial tilting effects of foundation.

Ground movement induced by triple stacked tunneling with different construction sequences

This study tried to explore the ground movement induced by triple stacked tunneling(TST) with different construction sequences. A case study in Tianjin, China was used to investigate the ground movement during the TST(upper tunneling(UT)). For this, a modified Peck formula was proposed to predict the surface settlement induced by TST. Next, three sets of finite element analyses(FEA) were used to compare the effects of construction sequences(i.e. UT, middle tunneling(MT), and lower tunneling(LT)) on vertical and lateral ground displacements. The results of Tianjin case and UT reveal that compared to a Gaussian distribution for a single tunnel, the surface settlement curve of triple stacked tunnels is a bimodal distribution. It seems that the proposed modified Peck formula can effectively predict the surface settlement induced by TST. The results of the three sets of FEA demonstrate that the construction sequence has a significant influence on the ground movement. Among the three construction sequences, the largest lateral displacement is observed in the MT and the smallest one in UT.The existing tunnel has an inhibitory effect on the vertical displacement. The maximum value of the lateral displacement occurs at the depth of the new tunnel in each construction sequence.

基于多域物理信息神经网络的复合地层隧道掘进地表沉降预测

复合地层中盾构掘进诱发地表沉降的准确预测是隧道工程安全建设与施工决策的关键问题。基于隧道施工诱发地层变形机制构建隧道收敛变形与掘进位置的联系,并将其耦合至深度神经网络(deep neural network,简称DNN)框架,建立了预测盾构掘进诱发地层变形的物理信息神经网络(physics-informed neural network,简称PINN)模型。针对隧道上覆多个地层的地质特征,提出了多域物理信息神经网络(multi-physics-informed neural network,简称MPINN)模型,实现了在统一的框架内对不同地层的物理信息分区域表达。结果表明:MPINN模型高度还原了有限差分法的计算结果,可以准确预测复合地层中隧道开挖诱发的地表沉降;由于融入了物理机制,MPINN模型对隧道施工诱发地表沉降的问题具有普适性,可应用于不同地质和几何条件下隧道诱发地表沉降的预测;基于工程实测数据,提出的MPINN模型准确预测了监测断面的地表沉降曲线,可为复合地层下盾构掘进过程中地表沉降的预测预警提供参考。

Study on Ground Deformation during Shield Tunnel Construction

Through the systematic analysis of the ground settlement generated by the process of shield tunneling,the relationships between ground deformation and construction parameters are studied in this paper.Based on the assumption of linear small deformation,a mathematical model of the relationship between ground deformation and construction parameters is set up.The principle and method of optimization for estimating ground deformation is studied.The actual measured data are compared with the results of theoretical analysis in a case.Considering different ground formations in different construction sites with different adverse effects on surface and underground structures,the ground surface deformations caused by shield tunneling is an aimed topic in this paper.The contributions and research implications are the revealed relationships between the ground deformation and the shield tunneling parameters during construction.

砂卵石地层矿山法隧道长管棚支护机理及应用

研究目的:砂卵石地层由于卵石强度高且颗粒分布无序,因此长管棚施工地层扰动大且支护效果难以保证,已成为制约隧道实际应用的一大瓶颈。本文以某轨道交通暗挖法车站为依托工程,采用理论分析、现场测试与数值模拟相结合的方法,对不同直径、不同环向间距条件下长管棚的支护效果进行了深入研究,以期揭示砂卵石地层矿山法隧道长管棚支护机理,提出最佳支护参数。研究结论:(1)砂卵石地层,暗挖隧道管棚超前支护条件下,地层沉降变形在向地表传递过程中,沿高度方向的折减率约为0.91 mm/m;(2)管棚超前支护效果对矿山法隧道工程安全具有决定性作用,须采取措施减小管棚施工扰动及相应地层变形;(3)根据单因素敏感性分析结果,地表沉降随管棚支护长度的增加而增大,随管棚环向间距的增大而增加,且基本呈正相关;(4)为有效控制地表沉降并提高工程经济效益,建议直径127 mm管棚环向间距宜为35 cm,支护长度宜为40 m;直径159 mm管棚建议环向间距宜为40 cm,支护长度宜为40 m;直径209 mm管棚建议环向间距宜为45 cm,支护长度宜为50 m;管棚直径与其最佳打设长度正相关;(5)本研究结论可为砂卵石地层矿山法隧道支护措施提供借鉴与参考。

基于迭代协同克里金反演的非均质地基固结沉降预测研究

目前地基固结沉降反演预测多是以土体均质或分层假设为前提,然而天然地基水文、力学参数具有空间变异性是客观事实。鉴于此,提出了一种基于迭代协同克里金的非均质地基土体参数反演方法,据此开展了利用沉降和超孔隙水压力观测数据反演非均质地基土参数的数值试验研究,并结合敏感度分析解释了不同类型观测数据对非均质地基参数反演解析度的影响机制。结果表明:此方法反演的参数场是最优无偏估计;同时采用沉降和超孔隙水压力观测数据反演刻画非均质地基比单独采用沉降或超孔压数据反演刻画的非均质地基解析度更高,用于地基固结沉降预测效果也更好;不同类型观测信息对不同参数反演刻画解析度高低与观测信息对参数敏感度数量级呈正相关关系。

墙顶支撑对开挖前抽水引发基坑变形控制效果研究

基坑内抽水可诱发基坑围挡向坑内偏转并进而导致坑外土体随动沉降;在基坑土方开挖前,由于基坑内的支撑体系尚不能被完整架设,仅坑内土体可起到约束围挡变形的作用,此时,坑内抽水引起的围挡偏转十分明显。本文系统研究了围挡顶部支撑对土方开挖前抽水引发基坑变形的控制效果,首先通过抽水试验探究了坑内外强水力连通时墙顶支撑对抽水引发基坑变形的影响,并建立了三维数值计算模型,研究了若干典型参数(如,坑内降水深度等)对围挡顶部支撑控制基坑变形效果的影响。研究发现,围挡顶部支撑仅能有效限制浅埋范围内的围挡侧移(本文约埋深11 m范围内),对深埋位置围挡侧移的限制效果十分有限,且随着降水深度的增加,变形控制效果越弱;对于开挖前抽水深度较大的工程,仅采用围挡顶部设置支撑的方法可能无法有效限制围挡侧移的发展;另外,围挡顶部支撑仅能有效限制基坑围挡后方一定范围内(本文约10 m)的地表沉降,且当坑内外有水力联系时,由于坑外水位下降是引起地面沉降的主导因素,围挡顶部支撑并不能明显体现限制坑外最大地面沉降的作用,此时,应结合坑外地下水回灌进行沉降控制。

河南地矿CORS省域地面沉降监测与机理分析

为了进一步研究省域尺度范围地面沉降,分析河南地矿CORS省域地面沉降的监测方法与其形成机理:针对河南省地面沉降分布特征,采用河南省地质信息连续采集运行系统(HNGICS)的62个基准站数据,时间序列为2012—2019年,获得国际地球参考框架下无净旋转条件的河南省三维速度场;然后分解三维速度场获得河南垂直速度场,发现河南省地面沉降呈现“西稳东沉”的现象,开封附近出现沉降漏斗;最后将河南省地质图、河南省耕地面积占比、河南省年均降雨量与沉降等值线对比分析。结果表明:不同地质条件下,河南省地面沉降分布于豫东地区的第四系冲积层,而相同地质条件下,耕地占比和区域年均降雨量是地表差异性沉降的主要原因,如豫东地区耕地约占总面积的70%,地下水用于农业灌溉是豫东平原出现沉降漏斗的主要原因。