Geo-Congress 2023: Geotechnical Data Analysis and Computation

In the present work, we carry out molecular dynamics (MD) simulations to investigate mechanism of crack in a single clay sheet. Uniaxial tension is applied to clay sheet in a strain-controlled manner. Crack initiates from the center of clay sheet and extends along the direction perpendicular to the loading direction. Finally crack extends through the whole slab and clay sheet is fractured into two halves. Bond breaking is first observed in silicon-oxygen bond, and later occurs in aluminum-oxygen bond. Numerical results show that crack propagation in single clay sheet is primarily contributed from bond breaking in Al-Ob and Si-Ob bonds. We find that crack length increases almost linearly with tensile strain. Atomic stress is calculated for both tetrahedral layer and octahedral layer considering the layered inhomogeneity of clay crystal structure. Stress intensity factor and energy release rate are computed based on MD simulation and are comparable to experimental results.

Most infinite slopes in tropical climates are frequently unsaturated in the beginning, and along with many soil properties like cohesion and friction, stability of these slopes is largely dependent on matric suction. These parameters are highly variable in the evaluation of slope stability for unsaturated soil. Hence, deterministic analysis is not enough to draw conclusions regarding stability of unsaturated slopes. This paper presents an analytical probabilistic method based on the theory of multivariate probability distribution to determine the stability of infinite unsaturated slopes. The stochastic soil parameters used in the study are cohesion, friction angle, unit weight, and matric suction. All the stochastic parameters were treated as uncertain variables which were defined by the normal distribution. The geometry-related parameters such as slope angle, slope height, and few stress variables such as net normal stress were regarded as deterministic parameters. The stability of the slope was determined by calculating the reliability index. Furthermore, the calculation formula for the probability density of slope safety factors was established. To inspect the correctness of the present methodology, the probability density of safety factors was determined by considering a conceptual infinite slope with arbitrary parameter values. Results showed that for the assumed set of parameters, the slope fails to meet the reliability index criteria of stability. A comparison between the outputs determined by the proposed technique and the Monte Carlo method indicates that the proposed methodology is efficient in order to calculate the stability of unsaturated slopes while taking the variability of various soil characteristics into account.

Rail track transitions are the discontinuities along the railway lines where conventional track adjoins the tunnels and bridges and are responsible for a sudden change in track structural properties. Frequent movement of fast passenger trains and heavy freights cause enhanced impact loads and differential settlements at these transitions that further lead to accelerated track degradation and increased maintenance costs. This paper presents a brief introduction to track transitions, the associated problems, and their causes and explains the occurrence of differential settlement at track transitions through analytical and numerical (2D FEM) modelling approaches. The stiffness effect on track settlement response under train loading is investigated, considering an Euler-Bernoulli beam on elastic foundation (BOEF). Vertical displacements of the track are calculated for various track stiffness values under train wheel loading (P = 10 tonnes). The results proved that the stiffer tracks undergo lesser settlements than the softer tracks. The effect of sudden stiffness variation is investigated by considering a one-step track transition where the stiffness suddenly changes from 80 MN/m (slab track) to 5 MN/m (ballast track). The results show that the settlement on ballasted track is almost 12 times greater than those on the slab track resulting in a large differential settlement at track transition.

Modeling granular column collapse provides an opportunity to compare simulated large deformation behavior with a variety of previous experimental and numerical findings to improve the understanding and characterization of newly developed soil models. Stemming from the popularity of modeling this experiment, the large deformation response of simple constitutive models, such as Mohr-Coulomb with strain softening (MCSS), is well characterized. In this study, the column collapse behavior of MCSS, NorSand, and bounding surface plasticity (BSP) models are compared. Simulation results indicate that constitutive model selection remains an important consideration. For sufficiently advanced models, such as NorSand, the selection of initial conditions has a greater effect on the runout behavior than the comparatively simple MCSS model. For the BSP model, a shorter runout is observed relative to both MCSS and NorSand models.

The paper demonstrates the use of the response surface method (RSM) to carry out probabilistic assessment of the bearing capacity of shallow footings seated on geosynthetic reinforced granular layers over randomly variable soft clay deposits. The method substantially reduces the number of Monte Carlo simulations required to carry out the probabilistic finite element limit analyses of the bearing capacity problem. A finite element limit analysis model based on the lower bound theorem is furnished and verified using some well-known analytical methods, and is then used to generate a large synthetic database of numerical results for bearing capacity of shallow foundations on a reinforced granular fill over randomly variably soft clay deposits. To this end, a permutation of the important influencing parameters is formed, and lower bound FELA-based limit loads are sought through optimization in MATLAB. A closed-form solution is formulated using RSM-based polynomials. The RSM equations, which are acquired from least squares regression analyses, are used to carry out probabilistic Monte Carlo simulations, and the results are presented in forms of cumulative distribution functions. Results from the probabilistic analyses are introduced into reliability-based design approach to render design loads for different reliability levels.

Coupling discrete element method (DEM) with computational fluid dynamics including Navier-Stokes theories (CFD-DEM) and Lattice Botlzmann method (LBM-DEM) has been used widely to model the response of soil foundation under increasing seepage flows; however, a comparison of these methods in predicting soil and fluid behaviours during fluidization has not been carried out in a rigorous way. The current paper will hence provide an evaluation on their performance by applying them to model a laboratory test where a sandy soil is subjected to fluidization process under increasing hydraulic gradient. A brief discussion about the differences in their theories and numerical algorithm is made before the DEM is used to simulate a representative soil element while NS-based CFD and LBM are used separately to model upward fluid flows. The mutual interactions between fluid and solid phases are carried out and update to each other through third-party platforms. The results show relatively similar hydraulic conductivity predicted by the two methods, which agrees well with the experimental data; however, the critical hydraulic gradient estimated by LBM-DEM coupling is found closer to the experimental value. The CFD-DEM coupling provides more stable computation through its averaged fluid variables, whereas LBM-DEM coupling can provide more detailed interactions between fluid and soil particles at micro-scale due to its high resolution. The study then suggests several conditions which can optimize the efficiency of using these methods in practical applications.

Landfill expansion is the best alternative to building new landfills due to the scarcity of land and a high rise in land cost. In the present study, the seismic stability analysis of the expanded landfill is carried out using the pseudo-static limit equilibrium method. As the municipal solid waste (MSW) landfill is highly heterogeneous, this study considered the variation of the unit weight of the landfill with depth under seismic conditions. Closed-form analyses are performed to determine the factor of safety under under-berm failure (failure surface passes under berm bottom). It is observed that with an increase in the height of the berm, the factor of safety also increased. With the rise in both horizontal and vertical seismic acceleration, the factor of safety decreases for under berm conditions.

NASA’s Artemis program aims to land the first woman and next men on the Moon while simultaneously working toward sustainable lunar exploration. A comprehensive understanding of the engineering properties and behavior of the lunar regolith is required for this program. As the amount of lunar regolith brought back to the earth is quite small, various lunar regolith simulants have been developed to simulate lunar regolith. In this work, LHS-1 Lunar Highlands Simulant, a high-fidelity, mineral-based simulant, is characterized for its physical and mechanical properties using conventional techniques such as sieve analysis, hydrometer analysis, and direct shear test. A discrete element method (DEM) model is then developed to simulate the grading-dependent behavior of LHS-1 in direct shear testing conditions. The DEM model is developed using the open-source software LIGGGHTS-INL and is calibrated and validated using characterization data of LHS-1. To the best of the knowledge of the authors, this is the first effort to develop DEM model for LHS-1. The validated DEM model can be used to simulate and predict the behavior of LHS-1 in complex systems such as the regolith-drill systems for more advanced applications.

In the past 15 years, the material point method (MPM) has been extensively used and developed to simulate large deformation problems in geomechanics, particularly landslides and other types of mass movements. The original formulation of the MPM, which assumes explicit dynamic time integration, often suffers from accuracy issues and numerical oscillation associated with its spatial and temporal discretization. On top of that, for fully undrained, i.e., incompressible, fluid-like flow slide problems, the numerical results obtained degrade significantly due to incompressibility-induced volumetric locking. At the same time, the incompressibility constraint also influences the allowable time step size for the explicit scheme, particularly since the numerical stability is strongly governed by the Courant–Friedrichs–Lewy condition. In the current work, a fully implicit MPM is implemented by incorporating the $\overline{\mathrm{B}}$ method to combat the volumetric locking issues. Higher-order basis functions are utilized to further improve the accuracy to alleviate errors associated with cell-crossing and numerical fracture. In the current work, the quadratic B-spline and the local maximum entropy basis functions are employed. Furthermore, to reduce the computational step associated with the rather expensive implicit time integration scheme and the higher-order basis function, a hybrid shared-distributed memory parallelization scheme is proposed by employing open-source domain decomposition and parallel linear solver software. The effectiveness and improvement of the proposed framework are qualitatively shown through simulations of progressive undrained failure of a 2D levee under gravitational loading.

This paper investigates the influence of sandy seabed characteristics on cyclic pull-out behavior of suction anchor for floating offshore wind turbine (FOWT) under cyclic wave loading on seabed and inclined environmental loads on padeye. The FE model in this study is developed by the combination of Biot’s theory, linear wave theory, and the elastoplastic constitutive soil model. Wave-seabed-foundation interaction (WSFI) is validated by the means of centrifuge test results, while the available data on cyclic direct simple shear (CDSS) test of sands is used to calibrate the CycLiqCPSP constitutive soil model. Results reveal that presence of non-liquefiable layer surrounding the skirt tip increases the stability of anchor remarkably. Moreover, parametric studies demonstrate that changes in the seabed’s sandy soil characteristics such as initial void ratio, shear modulus, and permeability can directly affect the further development of suction pressure beneath the anchor lid.

In a companion paper, we demonstrated the capability of a coupled discrete element method (DEM)–multi-body dynamics (MBD) framework in simulating self-burrowing behavior in granular media. In this paper, the same framework was calibrated by running direct shear tests and then used to systematically compare various horizontal self-burrowing strategies. The robot of interest has a minimalistic, modular design and mainly consists of a pair of cylinders with or without a cone or auger tip. The robot could achieve extension-contraction movement, and tips have options to rotate. A series of burrowing strategies were simulated: cyclic extension-contraction of the cylinders without a tip, with a static or rotating cone tip, or with a static or rotating auger tip. The rotation of the tip was only activated when the cylinders extended. It was found that without a tip, the kinematics is symmetric in time, and the robot does not have net translation with time. With a tip, the robots all burrowed horizontally due to the fact that the combinations of asynchronous extension-contraction and rotational motions break the kinematic symmetry. The burrowing speed is highest for the case with a rotating auger tip and is lowest for the case with a static cone tip. A net upward force (lift) was also found during the horizontal movement of the robot, which caused the robot to deviate from the planned trajectory and tend to move upward.

The importance of shear rupture in large-scale deformations of granular materials makes it a popular subject of many physical experiments supplemented by numerical analyses. Quite often the experiments provide the physical results on which numerical analyses rely for verification and validation. This workflow may not be reliable in applications centered around understanding the kinematics of individual grains within a developing shear zone, because the grains that are visible behind a transparent plexiglass boundary are simultaneously influenced by the presence of the boundary. This study critically evaluates this experimental-numerical workflow by quantifying the kinematics of grains adjacent to and away from the boundaries in simulations made directly analogous to their experimental counterpart. Base-offset experiments with a clean sand are performed multiple times to assess their average results and repeatability of results, and the results of numerical simulations are evaluated based on their similarity to the averaged experimental results. The level set discrete element method is utilized to capture the precise three-dimensional grain morphologies present in the experiments, and the sample preparation procedure of the experiments is replicated numerically. The development of the shear is analyzed in full quantitative detail in terms of grain displacements and rotations, and the transmission of forces throughout the experiment is deduced via grain-to-grain contact forces in the analogous simulations.

A three-dimensional (3D) nonlinear dynamic model of a pile-supported wharf created in FLAC3D was calibrated using data from a centrifuge test in prototype scale. Piles in the test and numerical model were subjected to the combined effects of inertial loads from the superstructure mass and kinematic loads from liquefaction-induced ground deformations during earthquakes. The recently developed pressure-dependent multi-yield surface (PDMY03) constitutive model was used to simulate undrained cyclic behavior of sands with different relative densities and the rockfill. The objective of the numerical modeling was to evaluate the capabilities and limitations of PDMY03 in capturing key dynamic responses of liquefied soils in 3D as well as soil–structure–foundation interaction between the wharf and the laterally spreading ground.

This paper evaluates the assumption of plain strain idealization of a strip footing rested on spatially varying random soil domain. The strip footing is numerically modeled herein as a two-dimensional (2D) and a three-dimensional (3D) problem using finite element method to get insights on the effect of the geometrical modeling of the strip footing on spatially varying soil domain. Cholesky decomposition based random field theory combined with Monte Carlo simulation is used to generate multiple random simulation of spatially varying soil domain. The simulated spatially varying soil domain is then coupled with the finite element analysis to perform the probabilistic analysis of stress and displacement fields due to the strip loading. The present study indicates that the geometrical modeling of the strip loading affects significantly the statistical parameters defining the uncertainties in the response of the strip loading due to soil spatial variability.

Traditional soil penetration activities such as site investigation and pile driving are energy-intensive and typically cause significant disturbance to surrounding soils and the environment. In recent years, researchers have attempted to solve the abovementioned problems by developing self-burrowing tools and probes on the basis of biological inspiration. Animals such as razor clams are capable of efficiently burrowing into soils using a dual-anchor mechanism, which cyclically alternates penetrating the back (shell) and front (foot) anchors to achieve forward movement. Due to inherent complexities of the self-burrowing mechanism, the interaction between a clam-inspired probe and soil is not fully understood yet. This study employs models based on the discrete element method (DEM) to investigate this soil-structure interaction problem. The soil sample is filled with a scaled discrete quartz sand analogue. The self-burrowing behavior of the probe is fully modeled using force-control motion, which enables realistic interaction with the surrounding soil. The strategy of tip oscillation is also employed to reduce soil penetration resistance. The simulation results show that the self-burrowing probe can achieve significant advancement, especially when tip oscillation is used. Micromechanical observations such as contact force networks and displacement fields are also provided to better understand the interaction between the soil and the bio-inspired probe. This study provides meaningful guidance for future self-burrowing probe development.

Guardrails, consisting of soil-embedded piles tied with a thrie-beam rail, are safety barriers installed along highways to shield errant motorists from roadside hazards and minimize the risk of serious injuries. Soil resistance plays a crucial role in the way that guardrails deflect when impacted by vehicles. Guardrails require special consideration over culverts as the soil fill depth provided over culverts is not generally sufficient to embed standard 1,016-mm (40-in.) deep guardrail posts. Various methods have been developed to enable post attachment to culvert structures. It would be desirable to use more closely spaced posts at a reduced embedment depth rather than introduce new hardware. However, the performance of the guardrails with post embedment depths less than 40 in. has not been evaluated. Thus, a need exists to determine the potential for using reduced post embedment and/or reduced post spacing with the guardrail to span low-fill culverts that satisfy MASH crashworthy criteria under TL-2 and TL-3 impact conditions. In this research, efforts were conducted to investigate the load-deflection behavior of a single post under impact through a series of dynamic, large-scale, component testing (with a surrogate vehicle, namely bogie) and numerical simulation using LS-DYNA. Baseline dynamic tests were performed on a steel W6x16 post embedded in crushed limestone soil with different depths impacted by a bogie vehicle. Additional six bogie tests were conducted on W6x8.5 posts with various depths. Test results were used to develop simplified numerical modeling of the soil-pile/soil-post system when impacted by a vehicle. The validated model would be used to study other impact cases with different vehicle models and post embedment depths compared to associated full-scale crash test data and investigate the accuracy of the proposed soil modeling method to represent soil and post behavior in crash tests.

This study aims to evaluate the load-settlement behavior of five different piled raft foundation case studies using two-dimensional (2D) and three-dimensional (3D) finite element analyses. Hydre Park Barracks in London and Messeturm, Torhaus, Westend, and the Skyper Tower in Frankfurt are five well-instrumented and monitored buildings with piled raft foundations. Their foundation systems were modeled via PLAXIS 2D and 3D software, and the computed load settlement behavior including maximum and differential settlement along with the pile loads and bending moments were compared with the field measurements. The results indicate that 3D models predict the load settlement behavior more accurately compared with 2D models, particularly on structures with a non-symmetrical pile layout. Furthermore, it is also concluded that while 3D finite element analysis yields more accurate results in expense of complexity of model setup and computational time, 2D models provide conservative results with simplicity and faster computation.

Furthering human exploration of the Moon will require the use of resources that may already be readily available on the lunar surface. This includes regolith and other materials that exist in varying depths depending on the location. However, accessing these materials and resources will require an accurate understanding of the geotechnical properties of the lunar regolith on the Moon’s surface, which can be determined through evaluation of the cone penetration process. This paper aims to develop a three-dimensional discrete element model (DEM) to gain an understanding of how lunar regolith reacts during cone penetration. To simulate the particles of lunar regolith, clumps are designed in a DEM using the calibrated contact parameters of the JSC-1A lunar regolith simulants. The DEM models are used to evaluate the regolith reactions such as contact force, velocity, displacement, and rotation of the regolith particles during the penetration process. The tip resistance and sleeve friction as they relate to an increase in depth are presented in this paper. Sensitivity studies are conducted to study the effects of different cone angles and cone diameters on the resulting tip resistance and sleeve friction during the penetration process.

Free-fall penetrometer (FFP) testing is a simple and rapid way of seabed sediment characterization that is more cost-effective than conventional geotechnical testing. Due to the dynamic nature of soil penetration during free-fall impact, loading rate effects are also recorded. As a result, a process to filter out the dynamic effects is required. However, removing the effects of high-strain rate (HSR) loading is challenging and favors the use of empirical correlations over analytical solutions. This study aims to numerically analyze FFP calibration chamber testing of saturated loose sands using the material point method (MPM). A fully coupled hydro-mechanical formulation is used to simulate the transient drainage that occurs during FFP deployment. Numerical models using material properties based on the bench-scale testing performed in quasi-static conditions result in lower tip resistance than the one obtained from the calibration chamber test. Parametric analyses show that dilation plays an essential role in the response of saturated sand and that HSR effects must be considered to capture sand behavior in such conditions. Finally, a comparative analysis of numerical CPT (quasi-static) and FFP test results is conducted to examine the effectiveness of logarithmic strain-rate correction procedure that is commonly used in practice.

A fully coupled three-dimensional (3D) thermo-hydro-mechanical (THM) model is developed for simulating the complex multiphysics process of permafrost thaw. The 3D formulation allows the analysis of thaw consolidation problems with complex geometry and boundary conditions. The thermal, hydraulic, and mechanical fields are coupled in this model. Governing equations are derived based on the laws of conservation of each field: conservation of energy for the thermal field, conservation of mass for the hydraulic field, and conservation of momentum for the mechanical field. Physical processes such as heat conduction, phase change, thermal convection, fluid flow due to pore water pressure, elevation, thermal gradients, and force equilibrium based on effective stress theory are considered in this model. The model is then applied to simulate the thaw consolidation of permafrost. The simulation results show that excess pore water pressure is generated in the soil during thawing. The soil then experiences a time-dependent settlement following the dissipation of excess pore water pressure. The results prove that the THM model adequately captures the thaw consolidation process of permafrost.

Shear stress history controls the pore pressure response in liquefiable soils. The excess pore pressure does not increase under cyclic loading when shear stress amplitude is lower than the peak prior amplitude—the shielding effect. Many sophisticated constitutive models fail to capture the shielding effect observed in the cyclic liquefaction experiments. We develop a data-driven machine learning model based on the LSTM neural network to capture the liquefaction response of soils under cyclic loading. The LSTM model is trained on 12 laboratory cyclic simple shear tests on Nevada sand in loose and dense conditions subjected to different cyclic simple shear loading conditions. The LSTM model features include the relative density of soil and the previous stress history to predict the pore water pressure response. The LSTM model successfully replicates the pore pressure response for three cyclic simple test results considering the shielding and density effects.

The performance evaluation of a potentially unstable slope involves two key components: the initiation of the slope failure and the post-failure runout. The finite element method (FEM) excels at modeling the initiation of instability but quickly loses accuracy in modeling large-deformation problems due to mesh distortion. Hence, the FEM is unable to accurately model post-failure slope runout. Hybrid Eulerian-Lagrangian methods, such as the material point method (MPM), offer a promising alternative for solving large-deformation problems, because particles can move freely across a background mesh, allowing for large deformation without computational issues. However, the use of moving material points in MPM for integration rather than the fixed Gauss points of the FEM reduces the accuracy of MPM in predicting stress distribution and thus failure initiation. We have created a hybrid method by initiating a failure simulation in FEM and subsequently transferring the coordinates, velocities, and stresses to MPM particles to model the runout behavior, combining the strength of both methods. We demonstrate the capability of the hybrid approach by simulating the collapse of a frictional granular column, comparing it to an empirical solution, and evaluating a suitable time to transfer from FEM to MPM by trialing multiple iterations with transfers at different stages of the collapse.

The goal of this study is to reveal the multiscale mechanism soils undergo during heating of retained backfill. Free-drainage soils surrounding energy piles or retained behind thermo-active tunnels or retaining walls experience heating/cooling due to heat exchange with the ground. Soils, therefore, expand/contract, rotate, displace, and change its stress state, which results in changing the lateral earth pressure experienced by piles/walls. To study the particulate mechanism of such stress state change, a two-dimensional (2D) discrete element model is utilized. The retaining wall is modeled numerically using a rigid wall, while the soil is represented by circular disks with varying particle sizes. The backfill was prepared by means of pluviation. Upon setting up the initial model, the temperature of the retaining wall is gradually changed, and the lateral earth pressure during different simulation stages is monitored. The results show that heating the modeled cohesionless backfill increases the at-rest lateral earth pressure. Finally, the evolution of the lateral earth pressure with temperature is justified by the heating-induced anisotropy within the heated backfill.

This study evaluates the behavior of diffused double layer (DDL) around cohesive clay particles under freezing temperatures using molecular dynamics. A better understanding of the physical molecular interactions related to the diffused double layer behavior between frozen soil and water is the key contributions of this study. The electrolyte system for this model is built by implementing the kaolinite system suggested by Bish combined with SPC/E water system in LAMMPS (large-scale atomic/molecular massively parallel simulator). Interactions between water and basal surfaces of kaolinite clay particles are studied due to its larger surface area compared to the clay edges. Simulations are proceeded under different temperatures ranging from 173 K (about –100°C) to 373 K (about 100°C) to capture the thermodynamics and diffused double layer behavior for the combined electrolyte system consisting of frozen soils and water. Quantitative analysis like radial distribution function is applied at the different temperatures to capture the variations in the behavior of diffused double layer. The results show a different pattern of thermodynamic properties but very similar radial distribution function analysis, which yields a minimal different behavior of diffused double layer for clay-water electrolyte system under freezing temperatures.

In this study, we present a multi-phase field thermo-hydro-mechanics model that simulates the growth and thawing of ice lenses and the resulting frost heave and thawing settlement in multiphase porous media. Based the Miller's theory, we assume that the ice lens can be viewed as a segregated ice inside the freezing induced fracture. In this model, the ice lens growth is implicitly captured via combination of two phase field variables that represent the state of the fluid and the fracture, respectively. Compared to phenomenological approaches that indirectly capture the freezing effects on the shear strength and stiffness, this approach distinctively captures the homogeneous freezing and ice lens growth by freezing characteristic function and the driving force, respectively. Evolution of two phase field variables are induced by their own driving forces that capture the physical mechanisms of ice-water phase transition and the crack growth, respectively, while their governing field equations are bundled with balance laws such that the coupling among solid deformation, fluid diffusion, heat transfer, phase change, and damage evolution can be observed numerically. 

Studying the behavior of surrounding soil in in situ environments during earth pressure balance (EPB) shield tunneling is deemed of great importance, and the discrete element method (DEM) has been commonly adopted for replicating the granular soil and capturing the qualitative behavior of ground response during tunnel advancement. The presence of fines has been proven to affect the soils’ skeleton structures and may dominate their shear strength. However, a significant body of research in tunneling engineering does not often consider sand-fine mixtures, although the natural soil consists of some portions of fines. This numerical study focused on simulating the EPB shield tunneling in granular material containing different fines contents, and the advancement of EPB shield tunneling in the longitudinal direction was reproduced using two-dimensional DEM models. The macroscopic behavior, such as horizontal stress at the tunnel face, and the surface displacement after excavation were investigated. In addition, the micro-mechanical aspect of soil behavior (i.e., contact density) was computed and examined. Finally, the effect of muck discharge rate, which was simulated by deleting the particles within the inlet area, was analyzed.

This study evaluates the influence of the inter-particle friction coefficient and particle elasticity on the critical state behavior of granular assembly using an open-sourced software, YADE. A range of μ between 0 and 0.5 has been used in the literature, with most studies with μ of 0.5. However, some studies suggested μ should be less than 0.5 to avoid an unusual increase in void ratio at the critical state with increasing mean effective stress (p′), while others did not observe that. This study evaluates whether the anomaly of the observation is due to the combined effect of interparticle friction and particle elasticity (E). A series of 90 DEM simulations have been performed using μ of 0.3 and 0.5 and E of 50 MPa and 100 MPa for both drained and undrained simulations up to 30% strain to observe the critical state behavior. The μ and E both influence the critical state behavior of the granular assemblies. Whereas the interparticle friction influences the vertical shifting of the CSL, the particle elasticity influences the horizontal shifting of the CSL with increasing mean effective stress. The larger value of μ increases the critical state void ratio (shift the CSL upwards), and low value of μ decreases the critical state void ratio (shift the CSL downwards) irrespective of the value of the E. The E, on the other hand, does not affect the CSL for low values of mean effective stress (p′<= 20 kPa) for granular assemblies with the same μ. However, for higher values of mean effective stress (p′> 20 kPa), the slope of the CSL horizontal increased for higher values of E of granular materials.

The failure of the Fundão Dam has been previously studied using two-dimensional approaches. The original investigation findings did not point to mechanisms of slope instability as the causes of the disaster. However, the complex shape of the dam and its concave and convex regions were not considered due to the limitations of the analysis tools employed. This paper presents two-dimensional (2D) and three-dimensional (3D) analyses that are more rigorous and better suited for the Fundão Dam geometry. Slope stability analyses were accomplished using the general limit equilibrium method (GLE). Critical slip surfaces were searched in 2D and 3D across the entire dam face. Drained shear strength parameters led to factors of safety (FS) that would indicate fairly stable conditions, with the lowest FS being 1.744. Values close to 1.0 were obtained considering undrained conditions. The 3D values of FS were very close to the 2D values, but with noticeably higher variations near the concave and convex regions. This leads to the conclusion that the dam shape near the setback region requires closer examination by means of 3D analyses, enabling the identification of critical zones.

For sustainable and environmentally friendly geotechnical engineering, biological soil treatment methods have recently been actively introduced. Biopolymer-based soil treatment (BPST) is regarded as a low carbon footprint ground improvement method with adequate strengthening and pore clogging characteristics. However, no detailed analysis of the relationship between the components that have a significant impact on BPST strengthening behavior has been carried out. Because BPST is frequently implemented as a mixture of soil, biopolymer, and water, different results can be obtained depending on the hydrogel phase and soil type. Using unconfined compressive strength (UCS) data for BPST obtained from laboratory test, support vector regression (SVR) was used in this study to predict the UCS and examine the dominant characteristics influencing the strengthening behavior of biopolymer-treated soil mixes.

Given that misalignment, while tunnelling in urban areas, can cause ground heaving or unfavourable soil deformations, it starts drawing great attention from scientists and engineers. In this study, the potential of applying deep learning (DL) and machine learning (ML) algorithms to earth pressure balance-tunnel boring machine (EPB-TBM) posture prediction and anomaly detection was explored. A predictive framework for the posture to a tunnelling project in loess region was proposed, including three phases: the exploratory data analysis to reduce dimensionality and highlight characteristics, the evaluation of DL-based models in posture prediction, and anomaly detection in posture deviations. The principal component analysis was used to decompose a tunnelling data set derived from Xi’an Metro Line 4, eliminating outliers and isolating its salient data features. The Pearson correlation coefficient was used to reduce irrelevant data features, while the partial dependence analysis aimed to marginalise other features and established the relationship between an individual character and the predicted target. After these EDA approaches, the feature-based subseries was normalised for training and examining the DL-based long short-term memory (LSTM) EPB-TBM posture predictors. Furthermore, several unsupervised anomaly detection methods were used to detect the posture prediction outliers. The validity of the proposed posture predictors was verified by comparing to the field measurements, and the results showed that the LSTM predictor performs well due to its complex nonlinear unit that constitutes larger deep neural networks to handle nonlinear problems. In addition, the deviation mechanism was revealed correspond to the relationship established by the exploratory data analysis and the geological conditions. These results highlight an exciting potential for the use of DL-based models to secure the sustainable development of surrounding environments from disturbing by tunnelling-induced misalignment.

The objective of this study is to develop data-driven predictive models for seismic energy dissipation of rocking shallow foundations during earthquake loading using decision tree-based ensemble machine learning algorithms and supervised learning technique. Data from a rocking foundation’s database consisting of dynamic base shaking experiments conducted on centrifuges and shaking tables have been used for the development of a base decision tree regression (DTR) model and four ensemble models: bagging, random forest, adaptive boosting, and gradient boosting. Based on k-fold cross-validation tests of models and mean absolute percentage errors in predictions, it is found that the overall average accuracy of all four ensemble models is improved by about 25%–37% when compared to base DTR model. Among the four ensemble models, gradient boosting and adaptive boosting models perform better than the other two models in terms of accuracy and variance in predictions for the problem considered.

This study demonstrates the application of a coupled discrete element method (DEM)–multi-body dynamics (MBD) framework in simulating the self-burrowing behavior of a robot in dry sand. In robotics, a robot can be modeled using MBD, in which each component of the robot is modeled as interconnected rigid or flexible bodies whose motions obey the laws of motion and are limited by kinematic constraints. In this study, DEM is coupled to MBD using Chrono—an open-source physics engine—to model the self-burrowing behavior of a two-auger robot in dry sand. The robot consists of a pair of horizontally aligned, auger-shaped bodies, which are connected with two rotational motors in MBD; and the sand is simulated as packing of monodispersed, frictional spheres in DEM. A typical co-simulation loop starts with the DEM module that solves the inter-particle and particle-structure forces and displacements; the particle-structure forces are then transferred to the MBD module to solve the dynamics of the robot; with the updated robot position and velocity, information is then transferred back to the DEM module. The rotation of the augers was set to different directions and speeds. It was found that the rotation of the two augers in the same direction enables horizontal translational movement of the robot, while rotation in the opposite directions hinders such movement: higher rotational speeds lead to higher translational velocities.

Due to the inherent variability of soil properties, geotechnical engineers often leverage multiple tools to estimate site subsurface conditions, including historical records, in situ measurements, and laboratory testing data. A comprehensive site investigation may be uneconomical for certain projects; thus engineers typically perform site characterization, which potentially could balance project costs and the extent of site investigation. To aid in preliminary analyses of site conditions, we have created a web-GIS database to compile soil engineering properties across the state of Oregon. The database, Georeferenced Oregon Soil Engineering Properties (GOSEP), is populated by existing data across projects from all over the state, including monotonic strength, laboratory characterization data, borehole logs, and in situ test results. In this paper, a detailed procedure showing kriging was used to generate three-dimensionally interpolated standard-penetration-test (SPT) data through the GOSEP interface. In addition, this study reveals the use of oblate spheroid as a kriging search geometry and the approach to dealing with modeling variograms with few data points. Ultimately, a cross-validation analysis was performed to study the accuracy of predicted values, and the limitations were discussed regarding the estimation.

Earthquake ground motion processing for next-generation attenuation (NGA) projects required human inspection to select high-pass corner frequencies (f_cHP), which is time-intensive and subjective. With growth in the number of recordings per event and interest in enhancing repeatability, we sought to develop automated procedures for f_cHP selection. These procedures consider signal-to-noise ratio (SNR) and non-physical features in the displacement time series that indicate high- and/or low-frequency noise effects. The procedures are implemented in a US Geological Survey software package (gmprocess). We extend previous procedures for SNR-based corner frequency selection to also check for low-frequency artifacts in the displacement record using a polynomial fit to improve f_cHP selection. We evaluate the performance of the SNR and polynomial fit criteria using recordings from the 2020 M5.1 North Carolina and the 2013 M4.7 Southern Ontario earthquakes. Data processed with the SNR-only criteria can have low f_cHP and displacement drift; the displacement check increases f_cHP and reduces displacement drift.

The vertical lift thickness of a soil that is placed, spread, and compacted during embankment construction is a critical parameter that ensures effective compactor energy penetration and a successful outcome of the compaction process. Consequently, most quality assurance/quality control guidance documents specify a maximum “not to exceed” lift thickness. In practice, it can be difficult to monitor lift thickness for large fill or embankment construction projects without frequent surveying, which adds cost and in some cases time delays to a project. One possible emerging approach for lift thickness monitoring is the use of UAV surveying. In this study, compacted lift thickness was inferred from unmanned aerial vehicle (UAV) imagery collected during active construction of an earthen embankment, which was processed using photogrammetric analysis techniques. To assess the accuracy of UAV surveying for lift thickness applications, 82 discrete locations along the embankment were surveyed utilizing traditional survey equipment. Physical lift thicknesses were determined at each of the 82 discrete locations utilizing a caliper and steel ruler. Spatial interpolation using isotropic kriging was performed on the lift thickness data set that was determined from photogrammetric analysis, to infer UAV-determined lift thickness at each of the 82 surveyed data points. The results from this study indicated that compacted soil lift thicknesses along the embankment ranged from 10 cm to 70 cm. When compared against physical lift thicknesses, lift thickness measurements made by the UAV in conjunction with isotropic kriging had a mean absolute error of 1.1 cm. The results from this study provide valuable insight on the relative accuracy of UAV surveying for lift thickness determination during active earthwork construction.

Machine learning algorithms can be used to detect landslides and develop landslide inventories following natural disasters. This study aims to evaluate the performance of different machine learning algorithms, specifically supervised methods, in detecting landslides on a complete high-quality landslide inventory. The M_w6.5 Lefkada earthquake that occurred on November 17, 2015, causing more than 700 landslides, was selected as the case study. The classification model used Worldview satellite imagery and digital elevation models (DEM) as input. Three machine learning algorithms were considered: random forest (RF), support vectors machine (SVM), and maximum likelihood (ML). All three machine learning algorithms classify landslides generally successfully, and RF achieves the best performance. Amalgamation is a typical issue in landslide detection and mapping, where multiple adjacent landslides are classified and mapped as one. This paper used flow direction to separate amalgamated landslides in the detection result. The result shows that the RF classifier captures large landslides more accurately than small landslides. Satellite imagery with higher resolution results in better classification performance than coarse-resolution imagery.

Sand boil associated with ground settlements is often observed due to soil liquefaction. A series of shaking table model tests were conducted to reproduce and examine the phenomenon. This study aims to utilize an adequate measuring tool, named the aerial triangulation technique in photogrammetry, to reconstruct the topography and measure the ground settlement at a top view location of the sandbox. A set of aluminum plates with known volume were used to verify the measuring technique. In addition, one 1-g shaking table test was conducted to prove that ground settlement and sand boil can be quantified. The volumes of the rigid plates, sand boil, and ground settlement were measured using a three-dimensional digital elevation model. The results show that the proposed image analysis can quantify objects’ volume on a model scale. In summary, aerial triangulation is a highly efficient and automatic technique that can provide complete spatial information.

Distributed fiber optic sensing (DFOS) offers the ability to create a continuous profile of strain and temperature measurements along an installed fiber optic cable. This technology has been successfully applied to a range of geotechnical applications, including recent advances in direct embedment in soil through installation in geotechnical boreholes. In this paper, the effectiveness of subsurface DFOS fiber optic strain and temperature monitoring to measure deformations and temperature changes is explored as a levee monitoring tool. The hypothesis is that vertical soil displacement and temperature changes in the subsurface can be measured and correlated with river water levels and uplift below the clay foundation. A field trial was performed to investigate this phenomenon, comprising of the installation of four vertical fiber optic arrays at a test site in Black Hawk, Louisiana, on a section of levee along the Mississippi River. A sand boil had previously developed in this area and was remediated using a combination of a gravel pad and a series of relief wells. Despite the remediation, the levee is still being actively monitored. A fiber optic monitoring system has been developed at U.C. Berkeley and deployed on site to provide readings of the long-term subsurface strain and temperature fluctuations along the landside of the levee. This data is supplemented by piezometric readings within the borehole arrays and monitoring of the river levels at a nearby station. The preliminary results have shown that concentrated tensile strains can be observed at the interface of the clay foundation and the underlying sand layer, correlating to fluctuations in the river level and foundation water pressures. This research demonstrates that DFOS can be used to detect small fluctuations in subsurface movements and temperature, as well as act as a tool for monitoring seepage and the potential for detecting issues prior to sand boil development.

Following the 2019 Ridgecrest (California) earthquake sequence, the Geotechnical Extreme Events Reconnaissance (GEER) association deployed and coordinated a reconnaissance effort that included teams funded by National Aeronautics and Space Administration (NASA), the US Geological Survey (USGS), the California Geological Survey (CGS), and the US Navy to document ground failure that had occurred at China Lake, Searles Lake, and surrounding areas. At Searles Lake, the teams found locations with ejecta and locations without surface manifestation, although the reconnaissance was relatively rapid, and most areas around the lake were not observed. Accordingly, two other data sources have been considered to develop a more complete spatial representation of ground failure: (1) damage proxy maps (DPMs) based on the analysis of multi-epoch synthetic aperture radar (SAR) data; and (2) optical (visible and near-infrared) satellite images. The Searles Lake lakebed lacks vegetation and is relatively level in elevation, making it an ideal location for using the remote sensing data. We begin by training a machine learning algorithm using observations from a small area to detect the presence of ejecta from the optical satellite images. Subsequent testing of the algorithm using observations from a broader area showed reasonably good results, but with larger rates of misidentification than in the training data. The algorithm in combination with the direct observations are used to generate post-event maps of surface manifestation at the same resolution as the satellite images, which are being used to validate the DPMs to facilitate future applications for rapid post-event ground failure detection and loss estimation.

Desiccation cracks occur in fine-grained soils due to evaporation and volumetric shrinkage. Characterizing soil cracking patterns can offer new insights into the key controlling factors and the underlying soil desiccation cracking mechanisms. However, the formation of desiccation cracks inherently involves a complex network problem, posing a challenging task for modeling and monitoring purposes. This study employs a deep learning technique to develop a new framework for characterizing soil cracking. We utilize a deep learning (DL) based convolutional neural network (CNN) architecture to segment and characterize soil desiccation crack images gathered from peer-reviewed publications. Further, we present a methodology to evaluate the pattern identification capability by considering the overall metrics (precision, recall, and dice) and details (the number of crack segments, crack total length, average crack width, and surface crack ratio) of the soil crack features. The results of the proposed framework are compared with those attained from a classical image processing technique of binarization by thresholding. The comparison shows the presented DL-based method offers significant improvements in the performance and accuracy of pattern recognition.

The US Army Corps of Engineers is developing and demonstrating the digital twin concept for flood risk management projects. Digital twins are being demonstrated beginning September 2022 at two flood control dams using existing technologies. A digital twin is a virtual representation of physical assets, processes, people, places, systems, and devices that provides both the elements and the dynamics of how the system operates and lives throughout its life cycle. Digital twins integrate data, artificial intelligence, machine learning, and software analytics with spatial networking to create living digital simulation models that update as their physical counterpart changes. A baseline framework was established for implementation of version 1.0 digital twins. Version 1.0 of the digital twins is focused on producing facility-specific decision-support information identified by dam facility personnel. These two systems will reveal best practices and result in information technology (IT) products that are scalable, expandable, and reusable for many other infrastructure types. The major components of this effort are the IT system framework, remote sensing and spatial data, user interface and visualization, and performance prediction and risk analytics. This paper describes the digital twin system design, the current system framework, the status of the efforts, lessons learned, and future research needs to effectively implement digital twins.

Degree of water saturation in soils has a key role in the response of geotechnical systems to earthquakes. In the absence of a remotely sensed tool to track soil saturation at a global scale, parameters such as annual average of water table depth, historical average of precipitation measurements, and topographic conditions are considered as proxies for soil saturation condition, the Soil Moisture Active Passive (SMAP), an earth observation satellite, has been launched in 2015 by National Aeronautics and Space Administration (NASA) to record soil moisture content in approximately the top 5 cm of the soil layer over the global land area with a three-day average temporal resolution. In addition, assimilating SMAP observations with estimates provided by land surface models, SMAP researchers can monitor root zone soil moisture in the top 1 m of the soil column. This paper shows how SMAP, as a high accuracy and near-real-time database, can be used to study the relationship between soil moisture and seismic ground response, which includes tracking the earth’s surface soil moisture variations caused by seismic events as well as flagging potentially damaged regions for earthquake reconnaissance teams.

Interferometric synthetic aperture radar (InSAR), light detection and ranging (lidar), and instrumentation have been tools in the geotechnical engineer’s toolbox for decades. However, these technologies have experienced improvements and cost reductions on vastly different timelines; therefore, engineers have generally employed these technologies separately for evaluations and long-term monitoring. Individually, these technologies provide unique benefits for a given project, but combined these technologies provide a holistic view of a geotechnical asset. Satellite InSAR systems, such as the European Space Agency’s Sentinel-1, provide a nearly temporal continuous picture of past and present movements in many discrete locations on and around an asset. Tripod-mounted lidar provides a spatially continuous 3D point cloud of assets at discrete points in time using the laser time-of-flight principle. Lastly, instrumentation, such as tiltmeters and shape arrays, provides automated continuous monitoring of movement on an asset at specific locations. Here, we present an example of a vertically integrated workflow using InSAR, lidar, and instrumentation for monitoring retaining wall performance. This study includes monitoring of several mechanically stabilized highway retaining walls in the southern region of the United States. The proposed workflow combines the three technologies to overcome their individual limitations and provide spatially and temporally continuous asset monitoring. For example, the long-standing challenge of locating a few high quality tiltmeters in the most significant areas of concern is partially answered by identifying wall locations that exhibit geometric anomalies in the lidar point clouds. Additionally, evaluating regional historical and ongoing InSAR trends and local automated instrumentation data helps inform the frequency of subsequent lidar scanning. When vertically integrated, these tools allow for the data-driven selection of the frequency and extent of future monitoring and inspections. This approach allows engineers to tailor the monitoring program to each asset’s performance and enables owners to allocate limited inspection and maintenance resources more efficiently.

Early warning (EW) can be defined broadly as the provision of timely advice prior to the occurrence of a highly unsafe event. Efficient data use for EW is a critical component of failure management in general, which encompasses activities such as vulnerable zone mapping, prediction, warning element, prevention, planning, and action. The objective of the current study was to propose an early warning protocol (EWP) for highway slopes (HWS) constructed on high-plastic clay in Mississippi. Six HWS were investigated, instrumented, monitored, and considered reference slopes. During this study, field investigation, lab testing, numerical modeling, and data analysis were implemented. Based on the analysis results, field observation, inventory, slope stability factor, and action plan were found to be the main components of the EWP. Moreover, it was found that the observe, plan, decide, and act are the key predecessors of the proposed EWP. This study developed and presented a simple, robust, and low-cost EWP for HWS failure mitigation that other state transportation agencies can widely deploy and operate.

Within the underground construction industry, transfer and storage of data is an important aspect of all projects. To increase the efficiency and accuracy of data transfer, DIGGS, a standardized format for structuring data that can be used to transfer data, has been developed. The original DIGGS schemas included geotechnical investigation and rock grouting project data, but efforts have been made to include all underground construction activities in DIGGS. This effort started with the introduction of soil permeation and compaction grouting data to DIGGS. The changes made to DIGGS under this effort successfully integrated soil grouting into DIGGS while also formatting the schemas for future modifications as other types of projects are added to DIGGS. This paper will discuss the procedures used to modify the DIGGS file format to include soil grouting and the changes made to the schemas, and demonstrate the effectiveness by applying DIGGS to real project data.

The very high level of geotechnical uncertainty and geotechnical risk during the planning and design phase of a tunnel or underground construction project is attributed to current site investigation techniques that result in a very low fraction (e.g., less than 0.01%) of the ground being characterized. This paper demonstrates how the extensive collection of real-time geotechnical monitoring and tunnel construction equipment data can be used to update local estimation of key geotechnical parameters that in turn enables the reduction of geotechnical risk. The paper first demonstrates how site investigation data can be used to develop pre-construction predictions of important subsurface project aspects, including tunneling-induced ground deformation and TBM performance prediction, e.g., tool wear, clogging risk. These apriori probabilistic predictions carry a high degree of uncertainty due to the very low percentage of ground that is characterized. The value of such predictions is therefore tempered by such high uncertainty. The uncertainty can be dramatically reduced through Bayesian updating, herein applied to tunneling-induced ground deformation.

We present a relational database of earthquake ground motion intensity measures and associated metadata for the state of California. NGA-West2 project spreadsheets have been adapted into a relational database format, and the data set has been expanded to include contributions from earthquakes, generally with magnitudes greater than M3.9, that have occurred since the conclusion of the data synthesis component of the NGA-West2 project in 2011. Aside from the newly added information, some site metadata fields have been updated for some ground motion stations. The relational-database is composed of multiple tables connected through a series of primary and foreign keys. We use various data types (beyond integers and floats) to increase the storage efficiency for several types of data. Currently the database includes 33,422 ground motions recorded at 2,739 stations for 478 events within or close to California. The database was designed to also accommodate the various fields included in other NGA databases, including NGA-East and NGA-Sub. This will eventually allow for certain database tables to be merged for all event types (e.g., a single site table could be created) in one central location.

ASCE Report Card of 2020 rated transportation infrastructure poorly for many states across the United States. Mississippi has its share of challenges regarding deteriorating multimodal infrastructure. Integral parts of the spatially distributed infrastructure are geotechnical assets such as highway embankments, retaining walls, subgrades, and bridge foundations. Traditionally, Transportation Asset Management (TAM) programs have primarily focused on pavement and bridge assessment with minimal attention to geotechnical assets. Geotechnical Asset Management (GAM) is still in the nascent design and implementation stage across the United States, including MS. There is a need for specific guidance involving modern techniques to evaluate geotechnical assets and prioritize maintenance funds to reduce risk to the infrastructure systems. Highway embankments are one of the most common and at-risk geotechnical assets in MS. The risks stem from high rainfall and cyclic weather patterns and are further compounded by expansive clay in many of the embankments in MS. This study investigated the performance of failed highway embankments across the central MS region using multiple approaches. Temporally spaced data was collected through remote sensing techniques such as light detection and ranging (LiDAR), geophysical investigations such as electrical resistivity, and in situ instrumentation such as an Inclinometer. 3D laser scanning was performed using the terrestrial LiDAR equipment Trimble X7. Dense point clouds data was acquired from a highway embankment that had previously experienced extreme surficial deformations and failure. Topographical surfaces were developed using the dense point cloud data, and the bare ground was extracted. Temporally spaced surface profiles of the failed and undamaged areas of the slopes were created for comparative analysis. The surface profiles and the subsurface resistivity images were matched against the inclinometer and instrumentation data collected across the same period. This comparative analysis provided valuable insights into surficial soil movement with varying weather patterns and subsurface moisture variations. Combining remote sensing and geophysical investigations will help develop a modern performance monitoring methodology for integration into the GAM framework. Furthermore, the results from this study promise a robust cross-platform assessment approach that can be expanded to other geotechnical assets and help streamline Geotechnical Asset Management.

Slope stability is highly sensitive to soil moisture variation, especially in the case of highway slopes of central MS, typically constructed on expansive clay. Traditional borehole monitoring techniques do not offer a complete picture of the soil moisture variation throughout the slopes. In contrast to high-altitude, satellite-based remote sensing data, uncrewed aerial vehicles (UAVs) with optical cameras and thermal sensors offer higher resolution, data-loaded pixels. Furthermore, UAV flights can be precisely timed to provide ambient sunlight exposure, and the freedom to acquire photos when desired is an added advantage. These advantages make the UAV-based imagery particularly suited for capturing reliable soil moisture variation data within the vadose zone of highway slopes. To this end, in this study, a UAV equipped with FLIR’s Zenmuse XT2 dual-sensor payload unit was used to capture temporally spaced optical and TIR images for a highway slope in central Mississippi (MS). The radiometric thermal infrared (TIR) images were post-processed, and pixel temperature values (°F) were obtained across the slope’s soil surface. The surface temperature readings were then correlated with ground truth soil moisture content (SMC). The ground truth SMC data were obtained through near-surface soil sampling and in situ moisture sensors in the vadose zone. A single-parameter regression model was developed to predict soil moisture from diurnal land surface temperature (LST) variations. Similarly, red, blue, and green (RGB) pixel values were extracted from the optical images and calibrated to obtain an RGB vs. SMC model with suitable fitting parameters. The results of this study are promising and set the stage to implement a time-efficient and non-contact site characterization method to capture soil moisture content variations within the vadose zone.

The rise of data capture and storage capabilities have led to greater data granularity and sharing of data sets in geotechnical earthquake engineering. This broader shift to big data requires ways to process and extract value from it and is aided by the progress in methodologies from the computer science domain and advancements in computer hardware capabilities. General machine learning (ML) models typically receive a set of input parameters and run them through an algorithm to gain outputs with no constraints on the parameters or algorithm process. Three topic areas of ML applications in geotechnical earthquake engineering are reviewed and summarized in this paper: seismic response, liquefaction triggering analysis, and performance-based assessments (lateral displacements and settlement analysis). The current progress of ML is summarized, while the challenges and potential in adopting such approaches are addressed.

In recent years, many researchers developed different methods for the estimation of liquefaction-induced lateral spreading displacements that caused major damage during earthquakes. While some of the studies are based on simplified analytical methods, some are based on artificial intelligence applications that are effective in solving many important and complicated problems. Within the scope of this study, an estimation method based on artificial intelligence is used to predict the liquefaction induced lateral spreading by using a comprehensive data set. For this purpose, a more up-to-date data set is created by adding the data of 2010 Darfield and 2011 Christchurch earthquakes to the existing data sets produced with standard penetration test data used in the past. Additionally, since majority of the prediction models consider the thickness of the liquefied layer, fine grain content, average grain diameter, earthquake magnitude, maximum ground acceleration, and distance to the seismic source as the most important parameters on liquefaction-induced lateral displacements, the data set used in this study is improved by involving the velocity-based intensity measures. As a result of the study, the relative importance of each selected velocity-based intensity measure on the lateral spreading displacements is examined and compared with each other in order to reveal their capabilities for the calculation of liquefaction-induced lateral displacements.

Reservoir embankments are built to retain water for various needs including water supply, flood damage reduction, flow channelization, and safeguarding lives and property. Due to the critical nature of these infrastructure assets, they are frequently monitored and maintained to ensure continued functionality and compliance with the safety guidelines. The routine traditional inspection practices are subjective to the inspector’s experience. The advancements in unmanned aerial vehicle (UAV) platforms and lightweight sensors have provided a solution to conduct proactive monitoring of these assets. A reservoir, located in restricted airspace, was inspected using a UAV equipped with an optical camera to conduct qualitative and quantitative condition assessments. Multiple flight missions were conducted at different flight altitudes to obtain high-resolution data of the embankment and the coverage of surrounding areas. Multiple 3D mapping products of the reservoir were built by processing the same set of aerial images collected using photogrammetric techniques. The stabilized soil embankment was monitored quantitatively to assess the distress conditions. Further, the visualization of the 3D models provided multiple views of the inlet pump, aeration systems, and overflow/emergency spillway and an immersive experience to the inspector, which could have been either prohibitively costly or physically impossible, or a combination of the two, to achieve using traditional inspection practices. The 3D model developed in this study will also be used as a comprehensive base map for timeline monitoring of these assets. This study will provide an idea about best practices for planning, collection, and processing of aerial imagery, which will be of interest to the agencies managing dam and levee assets.

This study aims to develop a multiphysics-enriched mixed reality game system that provides an integrated geotechnical learning experience with visualization, collaboration, and simulation tools. This paper summarizes the game development works. Essential game components, such as player characters, geotechnical testing devices and tools, and laboratory environment, were created for the virtual game. Multiple mini-games incorporating geotechnical concepts are designed to further encourage the players to play the game. The thermal conductivity and direct shear tests will be performed in a virtual walk-in environmental chamber with controllable temperature. Using these experimental results as modeling inputs, finite element simulations are integrated to support the geothermal pile analysis under field conditions. Players will be able to visualize the simulation results through the augmented reality technique. The entire game is created as a mobile game platform, which enables players to use personal smartphone or tablet devices. The work is expected to improve undergraduate students’ interests and success in geotechnical engineering coursework, as well as provide developmental insights and educational background to inform researchers who seek to develop similar games.

The Cretaceous black shales of the Great Plains of the United States are prone to slope instability and surface erosion problems. Fractured and unfractured shale materials have brittle mechanical behavior and rapidly weather into fine-grained soils. The shales are host to bentonite beds with liquid limits exceeding 200. Between the brittle nature of the shales and the weak bentonite beds, erosion and slope instability easily initiate. To help civil and geologic engineering students learn more about shales, erosion, and slope stability, a living laboratory has been constructed on the campus of the South Dakota School of Mines and Technology. This living laboratory is a poorly vegetated shale slope with ongoing erosion and slope instability problems. The site is used for research, and students are active participants in the research, with students sampling soils, developing experiments, surveying, and making instrument readings in several courses. Research includes erosion and instability rates, revegetation and erosion control methods, and use of automated monitoring systems. Living laboratories on campuses are an opportunity to couple teaching and research for an approachable combination that is effective at helping K-16 students understand fundamental concepts of geotechnical engineering.

Undergraduate geotechnical engineering courses are generally focused on theoretical aspects of soil mechanics and traditional design methods, most of which are based on limit equilibrium analysis. In current practice, however, thanks to advances in hardware, software, and soil material modeling, geotechnical engineers routinely rely on numerical modeling for analysis and design. To best prepare future engineers for the profession, exposure to computational methods and to software used in industry is therefore desirable at both the undergraduate and graduate levels. In addition, numerical modeling programs can be effectively used to illustrate fundamental concepts, to rapidly explore the impact of soil properties and design parameters, and to provide exposure to realistic geotechnical engineering design problems, thus enhancing students’ learning and classroom experience. This paper presents the experience of incorporating a finite element method (FEM) software used in industry into the undergraduate Geotechnical Engineering I course offered at Purdue University, through two laboratory modules built around key topics covered in the course. Feedback was collected from the students at the end of the semester through an anonymous survey designed to assess different aspects of student engagement (behavioral, emotional, cognitive, agentic). The survey results showed that the FEM labs improved student attention and promoted a deeper interaction among team members. Students also felt that the FE analyses were effective in bringing concepts to life compared to paper-based exercises, and appreciated the opportunity to work on a problem that had ties to an actual case history.

Geotechnical engineering requires the application of engineering judgment to successfully interpret field conditions and come to an appropriate decision. To practice engineering judgment, engineers evaluate their initial impressions when presented with new information and learn from past failures. This much is implicitly understood, but how engineers accrue engineering judgment is highly debated and is a research area still in its adolescence. Members of industry posit that engineering experience is essential for competent engineering practice, but traditional engineering education practices need to be updated with opportunities to acquire such experience. Internships help some students obtain relevant experience that entry-level engineers need. Unfortunately, not all students have the opportunity to experience hands-on internships, especially during the pandemic. Defining engineering judgment in the context of practicing engineering does not always map to a useful tool for educators to generate curricular designs that target development of this skill in the classroom. This prompted the need for a new paradigm regarding engineering judgment in undergraduate learning. Using a qualitative systematic review, we identified an emergent theoretical framework for engineering judgment that is comprised of seven emergent competencies. This paper presents the use of this emergent framework to design, implement, and assess a design project in a sophomore-level introductory geotechnical engineering course at a small private research institute in the Northeast of the United States. If we want students to learn the competencies included within engineering judgment, we need an explicit definition. If we want to support faculty as they respond to evolving ABET criteria, we need evidence of the use of these assessment tools for a range of student products. This paper presents the use of a research-based instructional strategy for utilizing student-centered goals, activities, products, and assessments (GAPA) aimed at developing engineering judgment in an undergraduate classroom.

Engineering judgment and critical thinking skills are widely sought-after skills in geotechnical engineering education in order to prepare students for engineering practice and to meet ABET accreditation outcomes. The Critical-Thinking Assessment Test (CAT) is an internationally recognized instrument for assessing critical thinking skills that has been developed at Tennessee Tech. The skills tested by the CAT include evaluation of information, creative thinking, learning and problem solving, and communication. In recent years, the Center for Assessment and Improvement of Learning at Tennessee Tech has begun to promote the development of CAT-Applications (CAT-Apps). CAT-Apps are sets of discipline-specific questions that are specifically designed to mimic at least one of the categories tested by the CAT. As students answer CAT-type questions within the context of their discipline, the students build critical thinking skills by exercising thought processes known to be linked with critical thinking. This paper describes the development and implementation of a set of 13 CAT-Apps targeted at entry-level geotechnical engineering. Some of the CAT-Apps were used within the context of such a course. Gains in critical thinking skills were measured by administering the conventional CAT test to a single class of students at the beginning and end of the course in which the CAT-Apps were used. The results are discussed, and recommendations are provided for implementing CAT-Apps or similar questioning techniques in other courses and contexts.

This paper’s objective is to document the incorporation of video technology with associated personality assessments into an undergraduate civil engineering curriculum for giving presentations. In 2020, video assignments were implemented into a laboratory that teaches stabilized soil, aggregates, concrete, and asphalt principles. This same laboratory has, for several years, required students to present to a panel of practitioners on open-ended topics based in the area of their laboratory experiments. A large practitioner and alumni survey documented the successfulness of these panels, which allow no electronics to be used, but also documented some items that technology based presentation assignments could improve. This paper summarizes the panel exercises and associated survey findings, and thereafter, describes the implementation of video presentations with associated personality assessments. The research question was whether video presentation assignments were useful in conjunction with the proven in-person practitioner panels toward improved student presentation skills. The collective findings of this paper support this laboratory incorporating video presentations, especially given the diversity in personalities identified during data collection.

Augmented and virtual reality (AR/VR) tools, once limited to the entertainment sector, entered slowly into the architecture, engineering, and construction (AEC) industries over the last decade. Additionally, wearable products, such as Microsoft’s HoloLens 2, have led to a new category of technology: mixed reality (MR). Together, these powerful technologies can immerse project stakeholders into a 3D data-rich digital environment to rapidly convey spatial relationships of project data such as subsurface conditions and location of underground utilities. Despite potential applications of these technologies and recent reductions in hardware and software costs, AR/VR/MR tools have been largely limited to tradeshow marketing and architectural presentations. Although these technologies have been successfully employed on many large infrastructure projects, several technological barriers and social trends have limited adoption within the geotechnical industry. For example, most commercial AR/VR/MR software generally lacks an automated pipeline from engineering design documents to the digital model, preventing real-time communication between designers and other project stakeholders. Additionally, the AEC industry has largely focused on using expensive and limited-purpose hardware, rather than utilizing the full potential of devices that stakeholders already own, such as smartphones and tablets. Here, we present several use cases where AR/VR/MR may improve engineering cognition on typical medium-sized projects. These cases illuminate some opportunities for quick and easy adoption of these technologies as well as highlight key barriers to use.