Challenges and Innovations in Geomechanics

We present numerical methods to simulate the propagation of discrete fractures embedded in a damaged zone. Continuum Damage Mechanics (CDM) models are implemented in a Finite Element (FE) analysis code. A damage threshold defines the beginning of micro-crack coalescence, when a discrete cohesive segment opens. First, Cohesive Zone (CZ) elements are inserted between volume elements. The length and orientation of the discrete fracture are controlled by the magnitude of the energy released at integration points. The fracture path highly depends on space discretization, but the damage threshold is calculated automatically upon CDM model calibration. Second, an eXtended Finite Element Method (XFEM) approach is proposed. The fracture path is a function of the weighted average direction of maximum damage. Fracture growth depends on a nonlocal internal length, and the damage threshold is set empirically. Both coupling methods perform satisfactorily for simulating pure mode I or pure mode II fracture propagation resulting from micro-crack coalescence. However, the derivation of the Jacobian matrix in the XFEM makes it impractical to couple CDM and CZ models when the damaged stiffness cannot be expressed explicitly. For mixed mode fracture propagation and bifurcation problems, CZ insertion shows great promise, but mesh dependency and computational cost remain challenging.

Instabilities can occur in solids and fluid infiltrated granular systems affected by elastic, plastic and creep strains, friction, adhesion, sliding, rotation of grains/particles and surface forces due to readjustment or reorganization of the microstructure. Instabilities can occur at multiple locations at different states during deformation, e.g. transition from compactive to dilative volume change, peak condition and critical states at which failure or liquefaction initiates leading to final liquefaction. Conventionally, liquefaction is often identified based on the comparison of induced pore water pressure and the initial effective stress, which is considered to be an external method to represent the internal mechanism in the deforming material, which may not be realistic. In contrast, the disturbed state concept (DSC) and energy approaches can provide fundamental procedures and allow for internal mechanisms. The DSC provides a unique and basic model for the initiation and final liquefactions corresponding to the critical (Dc) and final (Df) disturbances, respectively. The issue like shear band formation occurs as a special case and can represent only one state of such distributed mechanisms. The DSC for liquefaction analysis is emphasized in this paper. A number of laboratories simulated and field validations for geomechanical problems are also presented.

Current computational platforms allow unprecedented opportunities for conducting seismic soil-structure interaction simulations. In geomechanics, capabilities such as coupled solid–fluid formulations and incremental-plasticity approaches allow for representation of the involved seismic response. Recent research that facilitated such endeavors in terms of response of ground-foundation-structure systems is presented. Representative numerical results are shown for a number of large-scale soil-structure systems. Graphical user-interfaces for conducting routine three-dimensional (3D) simulations are discussed, as an important element in support of wider adoption and practical implementation. In this context, Performance-Based Earthquake Engineering (PBEE) analysis of bridge-ground systems is highlighted as topical application.

This paper presents the development of self-developed cellular automata software for engineering rockmass fracturing processes (CASRock). CASRock is a combination of cellular automaton, rock mechanics, engineering geology, elasto-plastic mechanics and fracture mechanics. CASRock was initially developed to simulate the failure process of heterogeneous rock, and it is improved to focus on the fracturing process of engineering rockmass. The use of a cellular automaton allows CASRock to accurately represent the real fracturing characteristics of engineering rockmass. This process is a ‘down-top’ way handle the fracturing behaviour of rockmass. CASRock contains several mechanical models, including the 3D elastic-plastic-ductile-brittle model, for different applications in rockmass engineering. New stability analysis indexes, e.g., the RFD (rock fracturing degree), are incorporated into CASRock to evaluate the fracturing degree of engineering rockmass. A parallel cellular automaton updating rule is developed for large-scale stability analysis of engineering rockmass. The theory and software functions are briefly introduced. Several applications are presented to show the ability of CASRock to simulate engineering rockmass failure processes under complex conditions.

Quality of life is strongly correlated with power consumption. The geo-disciplines have a crucial role to play in the energy challenge by contributing solutions to all kind of energy resources from resource recovery to energy and waste storage. Energy geoengineering requires a broad understanding of physical processes (sediments, fractured rocks and complex multiphase fluids), coupled phenomena, constitutive models for extreme conditions, and wide-ranging spatial and time scales. Numerical methods are critical for the analysis, design, and optimal operation of energy geosystems under both short and long-term conditions. Furthermore, they allow “numerical experiments” at temporal and spatial scales that are unattainable in the laboratory. Yet, computer power can provide a false sense of reality and unjustified confidence; simulations face uncertainties related to the validation of complex multi-physics codes, limited data, excessive numbers of degrees of freedom, ill-conditioning, and uncertain model parameters. Dimensional analyses help identify the governing processes and allow for simpler and more reliable simulations. Educational programs must evolve to address the knowledge needs in energy geoscience and engineering.

In recent years, there has been an increasing amount of research on the Material Point Method (MPM) for modeling multi-phase coupled problems. Applying MPM in hydro-mechanical problems that interest geotechnical engineers have been explored in many of these studies . The explicit MPM method has been favored in dynamic large deformation problems due to its computational efficiency. However, numerically generated pore pressure oscillation has been a major issue. This paper presents a new formulation of MPM to model coupled soil deformation and pore fluid flow problems. The formulation is presented within the mixture theory framework, and pore water pressure is solved implicitly using a splitting algorithm based on the Chorins projection method. The splitting algorithm helps mitigate numerical instabilities at the incompressibility limit when equal-order interpolation functions are used. It reduces pressure oscillations and a time step size, which is independent of fluid compressibility and soil permeability. The proposed method is validated by comparing the numerical results with the closed form solutions of one dimensional and plane strain problems.

Strain localization zones in the form of shear bands or compaction bands in geomaterials are observed across scales from sub-millimetric (grain size) to kilometric scale (geological structures). Triggering and evolution of such narrow zones of localized deformation depend on many factors. The mechanical behavior of geomaterials is central for the formation of such zones. However, thermal, pore-pressure and chemical effects play a crucial role in shear and compaction banding. Moreover, the inherent heterogeneous microstructure of geomaterials plays a significant role during strain localization. As for faults, compaction bands significantly influence the stress field and fluid transport. In this paper, we shall review some recent advances in experimental testing and numerical modelling on strain localization in geomaterials. The effect of grain crushing as observed in deformation bands under high confinement can be introduced by combining Breakage models and higher order continuum theories leading to a new framework of constitutive models with evolving microstructure. A major difficulty of these models is to establish reliable methods for the calibration of the higher order parameters (such as the internal length) in laboratory experiments. An example of a direct calibration of these parameters based on Digital Image Correlation of images provided by X-Ray tomography is proposed for the study of compaction banding in a carbonate rock.

The thin layer of ice covering the polar oceans is a complex geomaterial that is constantly breaking and moving under the action of the wind and ocean currents and that experiences transitions between a brittle solid and a granular fluid state. We have developed a simple continuum mechanical framework in the view of representing accurately its dynamical behavior in regional and global sea ice or coupled climate models. It combines the concepts of elastic memory, progressive damage and viscous-like relaxation of stresses. Here, we present this framework and its ongoing numerical development based on Finite Element, Discontinuous Galerkin methods.

The hybrid finite/discrete element method (FDEM) is an innovative numerical technology that combines the advantages of large-strain finite elements with those of discrete elements, allowing for explicit modelling of the propagation of cracks in heterogeneous brittle rocks, as well as the interaction between their different constituent phases and materials. By modelling these physical interactions at the micro-scale, FDEM is able to reproduce the responses of excavation in layered and bedded rock at multiple scales as an emergent property of the model. Following an overview of the FDEM basic principles, failure process observed in shale rock will be presented and discussed within the manuscript.

This lecture examines the progress made in the last few decades concerning the use of geosynthetics for aiding in the construction on soft soils. Consideration is given to different soft soils ranging from peat to rate-sensitive soft clay and silt. Both relatively elastic and rate-sensitive reinforcements are examined. Consideration is given to basal reinforcement, prefabricated vertical drains, and embankments with reinforcement and other supports such as piles. Particular emphasis is placed on advances since the senior author’s 2002 Giroud lecture.

The simulation of cone penetration tests (CPT) poses still a challenge to numerical modelling because large deformations and large displacements have to be considered. In this work the Particle Finite Element Method code G-PFEM, which employs an updated Lagrangian description, is utilized. The use of linear elements in combination with a stabilized mixed formulation and frequent remeshing of critical regions ensures computational efficiency. The well-known Clay and Sand Model (CASM), which is a model based on critical state soil mechanics principles, has been implemented in G-PFEM and extended to account for effects of bonding and destructuration. Cone penetration in a low permeable silt under undrained conditions has been simulated and the influence of the degree of bonding and the rate of debonding on calculated cone resistance is evaluated. In an additional study the influence of the shape of the yield surface, which can be controlled by two input parameters in the model, on cone resistance is investigated.

The estimation of the Thermo-Hydro-Mechanical (THM) properties of heterogeneous rocks is influenced by the scale of the heterogeneity in relation to the dimensions of the test specimen, which will enable the testing of a suitable representative volume element. This places constraints on the testing techniques and the problem can be compounded when the rock has low permeability that presents challenges for saturation of the rock and required time for conducting basic tests to evaluate the deformability and transport properties relevant to fluid flow and heat conduction. The paper proposes the use of multiphasic techniques estimating THM properties of the heterogeneous Cobourg limestone.

The macroscopic response of geomaterials is controlled by the processes occurring at the microscale. Understanding these processes is key to interpret experimental data, understand fundamental modes of stress-strain behaviour, inform ‘continuum’ macroscopic constitutive models, and develop quantitative predictive tools based on Discrete Element Method (DEM) approaches. Unlike granular materials, mechanisms at the particle scale controlling macro-mechanical behaviour of clays are still largely ignored. This paper presents an analysis of elementary mechanisms of clay particle interactions with the aim of gaining an insight into behaviour of clay and advance the process of defining suitable contact laws to be implemented into DEM formulations.

Pile driving induced vibrations creates a huge concern for the construction industry since it may cause damage to the surrounding structures or settlement of the soil, depending on the intensity of ground shaking or vibration. It is essential to estimate the transmitted vibration intensity to avoid structural damages, which are highly dependent on the physical properties of the pile and the soil that acts as the transmitting medium. One widely adopted solution for the screening of pile-driving induced ground vibration is the use of infill trench in the path of wave propagation. The high damping and energy absorption capacity of rubber is well established in the past, making it an ideal material in vibration mitigation studies. In this paper, the use of sand-rubber mixture (SRM) as a trench infill material has been investigated to understand its effect on reducing the response of impact pile induced vibrations to adjacent structures. The SRM-filled trench barrier was provided as a passive isolation system, and field tests were conducted to evaluate its vibration screening performance under impact loading due to piling. The outputs from the analysis, in the form of acceleration-time history and Peak Particle Velocity (PPV), was obtained on either side of the trench with and without SRM infill. Overall, SRM was found to have a better performance with regards to the attenuation of surface waves. Further, a considerable reduction in peak acceleration and PPV was noted due to the introduction of SRM-filled trench barriers.

In view of its environmental and economic benefits, the use of Recycled Asphalt Shingles (RAS) and Reclaimed Asphalt Pavement (RAP) in Hot-Mix Asphalt (HMA) is on the rise. However, concerns over premature cracking and reduced fatigue life have limited their use in HMA. Polymer-modified asphalt binder is known to enhance the fatigue performance and resistance of HMA to cracking. This study was undertaken to evaluate the effectiveness of using a polymer-modified asphalt binder in asphalt mixes containing RAS and RAP to reduce fatigue cracking. For this purpose, two sets of HMA mixes, one set with a non-polymer-modified and the other set with a polymer-modified asphalt binder, both having identical gradations containing equal amounts of RAS and RAP were designed and tested for fatigue using a four-point bending beam method. It was found that polymer-modified asphalt binder effectively improved fatigue life of HMA mixes containing RAS, RAP, and their combinations. Also, use of RAP and a combination of RAS and RAP in asphalt mixes containing both non-polymer-modified and polymer-modified asphalt binders improved their fatigue lives. However, using only RAS resulted in an adverse effect relative to fatigue performance.

Grain crushing is a phenomenon occurring at the grain-scale along areas of stress or strain concentration where the stress imposed exceeds particle strength. The grain fragmentation determines fine generation and a local densification of the material, which may induce contractions, settlements, decrease of hydraulic conductivity, etc. This work reports the results of a wide experimental program aimed to link the macroscopic behavior of granular materials subjected to oedometric compressions to grain crushing accounting for the micro-structural features of the single grains, e.g. size and shape. The grain scale was assessed by the use of the X-ray tomography facilities to characterize the material and quantify what is occurring along deformation. Tests performed included one-dimensional compression tests on dry samples till high pressure (~50 MPa) in order to observe significant amount of crushing. The material adopted is the Light Expanded Clay Aggregate (LECA), an artificial granular material characterized by light, porous and crushable grains. The compressibility curves has been linked to the grain crushing phenomena upon loading. The beginning of comminution is ruled by the strength of the single particle and by the initial assortment of size and shape. As the micro-structure influences the grain crushing and the material response, vice versa the grain crushing influences particles size and shape. The collected experimental evidence is finally used to develop a simple constitutive model able to predict the evolution of material porosity with loading.

The suitability of LS-DEM to simulate drained and undrained cyclic behavior of sand is evaluated in this paper. A digital twin of a small specimen of the Hostun sand is used to simulate the cyclic average (macroscopic) response along idealized stress and strain paths, representing oedometer, drained and undrained triaxial conditions. The results are here only qualitatively evaluated based on comparison with real test data of different sands, since a) cyclic laboratory tests have not been performed on the actual specimen, and b) relatively low contact stiffness was used in this preliminary analysis in order to reduce the computational cost. The oedometer curves during virgin loading, unloading and re-loading and the cyclic secant stiffnesses during drained cyclic loading were captured qualitatively, but were significantly softer than the experimental ones. From the simulation of a drained cyclic triaxial test the development of accumulated volumetric strain seems reasonable, and the computed response during a cyclic constant volume (undrained) triaxial test looks very promising. Based on this initial study, more detailed LS-DEM simulations will therefore be carried out to improve the understanding of cyclic behavior of saturated sand.

In many instances, saturated cohesive soils are subjected to various types of cyclic/dynamic loading conditions such as earthquake loading, blast loading, oceanic wave storms, machine vibrations and traffic loading, etc. All of these loadings are different based on the rate and magnitude of loading cycles. Hence, the aim of the current study is to evaluate the undrained cyclic shear behaviour of saturated Nagpur soil subjected to various loading conditions. A series of strain-controlled cyclic simple shear tests were carried out on the compacted specimens of Nagpur soil at varying shear strain amplitudes (γs = 0.25, 0.5, 1.5, 2.5, 3.75, and 5%) and loading frequencies (f = 0.1, 0.5, 1, and 2 Hz). The hysteresis response was analysed in terms of shear modulus and damping ratio variation along with cyclic stiffness degradation and cumulative strain energy dissipation. Shear modulus and damping ratio were found highest for specimens loaded at the highest frequency with the lowest excess pore pressure generation. Rate and magnitude of stiffness degradation were reduced with a decrease in shear strain amplitude. Dissipated cumulative strain energy was observed to be maximum for higher shear strain amplitude and rate of loading. The results indicated that the frequency and amplitude variation greatly influenced the hysteresis response of saturated Nagpur soil.

To study the issue of moisture migration and density changes in the sample during the test, the authors conducted experimental studies on the changes in the clay soils physical and mechanical characteristics under triaxial compression conditions in various devices: cubic and cylindrical. Experimental studies conducted have made it possible to establish that the duration of the loading stage significantly affects the volume of the fluid being moved. It was established that regardless of the sample regime loading and the test equipment type, a complex stress state is formed in the sample, which contributes to the formation of different density and humidity zones. The dimensions of these zones, the soil density and humidity, its mechanical characteristics within the specified zones are not constant and change during loading. In the soil sample loading process, two mutually compensating processes occur. The first softening process is due to the formation of microcracks and macrocracks. Secondly, this is the process of hardening due to the restoration of water-colloidal bonds between the clay soil particles and the “healing” of defects. The hardening and softening of clay soil mechanism is established.

Infiltration characteristics of an unsaturated soil are of interest in both seepage and stability analyses and should be evaluated carefully. In this study, a new large-scale one-dimensional infiltration column test setup was developed, with an inner diameter of 0.24 m and a length of 1.3 m. In the tests conducted, 1.0 m of the column height was filled with soil and 0.3 m remained above for constant water head. The column was instrumented with five pairs of in-situ volumetric water content sensors and suction sensors. This paper explains the methodology used in the construction of the test setup and how the unsaturated properties were calculated for the tested soil, namely the soil water characteristic curve (SWCC) and soil permeability function (SPF). Infiltration tests were performed on a fabricated homogeneous clayey sandy silt similar to naturally available materials representative for Norwegian conditions. Soil specific SWCCs were established under steady state boundary conditions using the sensor outputs, and the results are presented. The instantaneous profile and wetting front advance methods, and relationships based on the SWCC were utilized to establish SPFs, and the results are discussed. A sensitivity analysis was run on the SWCC curve fitting parameters and effects of the parameters on infiltration time are presented. The results from the combined experimental and numerical analysis show that it may be possible to use the new test setup to develop unsaturated soil relationships, but accuracy and measurement range of the sensors are crucial to obtaining consistent results.

Interpretation of CPTU in silts is challenging compared to sands or clays since penetration occurs allowing partial drainage. Further insight into the soil response around a penetrating cone in silts is needed, including volumetric changes and pore pressures. This paper presents measured cone resistance, soil stress and pore pressure in silt during model scale penetration tests with penetration velocities varied by three orders of magnitude. The experimental results are back-calculated by finite element simulations to identify key aspects of the silt behavior during penetration of the cone.

Performance of asphalt pavements depends on the quality of compaction achieved during construction. Asphalt cores as an indicator of construction quality are not reliable because they typically cover less than 1% of the constructed pavement. Intelligent compaction (IC) estimates the level of compaction of the entire pavement layer during construction. IC rollers are equipped with accelerometers for measuring vibration, a GPS for monitoring spatial location, a temperature sensor for measuring surface temperature and an on-board computer for real-time execution of software and data storage. Although IC shows great promise as a quality control tool, there are concerns regarding the quality and analysis of data including missing data, data accuracy, data filtering, and data interpretation. Also, verifying compliance of the IC output with Department of Transportation (DOT) requirements needs accurate project boundaries. Project boundaries obtained from the onboard GPS may not be adequate for verifying compliance. In this study, the IC data from three pilot projects in Oklahoma were analyzed using the VETA (v 5.1) software, which is a map-based tool for viewing and analyzing IC data. Three different IC providers were used in collecting these IC data. A high degree of variability in collected data was observed, including inconsistent file naming, unspecified target for number of roller passes and inadequate layout of project boundaries. Despite variability, coverage, number of roller pass, compaction temperature, roller speed and roller frequency were found useful as indicators of compaction quality. Project size and operator training were also found to be important factors for successful implementation of intelligent compaction as a quality control tool.

Geotechnical parameters scatter in a wide range. On the one hand, this is due to the spatial variability of the subsoil, but also results of laboratory tests on reconstituted specimens of one sample scatter, as a completely homogeneous, reproducible specimen preparation is not feasible. For calculations according to the standards, characteristic shear parameters should be chosen as cautious estimate of the mean values. How this cautious estimate should be determined is not defined and therefore subjective. Often the results of shear tests are used as basis for the decision. In this paper, results of drained triaxial compression tests on a reconstituted, natural, widely graded soil are investigated. The specimens were prepared at same mean density but the results of the shear experiments scatter. The deviation of e.g. the peak strength is apparent. For the derivation of the Mohr-Coulomb parameter friction angle and cohesion according to the standards 3 or more stress levels have to be considered. The influence of the number of stress levels taken for the evaluation of the shear strength parameters is quantified. The evaluation of only three stress levels leads to a relatively large range of possible shear parameters. The two shear parameters friction angle and cohesion are statistically dependent - since they are two parameters of a linear regression. Therefore, they should be considered together. The scattering in the peak strength is probably caused by an inhomogeneous specimen construction. The influence of an inhomogeneous specimen preparation on the peak strength is investigated and proven in numerical simulations.

There are several well-established jacking force empirical models available for predicting the jacking force. However, the jacking force calculated by different empirical formulas have big difference and the face resistance is difficult to measure directly. These factors constrain the prediction of the jacking force in design stage. This paper introduces a pipe jacking technology with non-over excavation. The data of jacking force and steel pipe stress were recorded and analyzed in the process of pipe jacking. According to field monitoring data and the finite element numerical simulation results, the face resistance and the friction resistance can be obtained by back analysis. And the face resistance obtained is compared with the empirical value. The results show that JMTA and P-K models have practical significance in calculating face resistance F0, while Ma Baosong model underestimates the value of face resistance in silty sand formation. This research can provide a theoretical basis for the selection of jacking force prediction model.

To explore the difference between the critical state lines (CSL) of crushable and uncrushable sand with the same gradation as crushable sand at the critical state, two sets of DEM simulations are conducted with different confining pressure. A series of Large-scale triaxial compression test were carried out on rockfill material as the basis for calibration of discrete element model parameters. At first, a crushable rockfill sample was sheared to reach critical state. After that, the sample with the critical state gradation of first set was recovered to isotropic consolidation state with the same confining pressure and sheared to critical state under the same stress path. A crushable DEM sand sample was established with validated particle failure criterion and fragment replacement method. The DEM model of first sets were validated and calibrate by experiment results by the stress-strain behavior. And the breakage extent of DEM model was also investigated. The same parameters were used in the rest simulations. The stress-strain behavior of uncrushable sample and CSL of both samples were obtained. From the results, the important influence of the process of the grain crushing on the CSL location was observed although this two set of simulations have the same final PSD.

It is well established that the principal stress rotation (PSR) can lead to plastic deformations in soil, and soil is subjected to two dimensional shears in many geotechnical applications, which lead to two dimensional principal stress rotations. This paper aims to study impacts of two dimensional PSRs on soil behavior by using the DEM and experimentation, in which glass beads are used. In the two directional simple shear apparatus, the first shear is exerted on samples along one direction, followed by the second shear at varying angles to the first shear. The experimental results indicate that the angle between the first and second shears have a great impact on the sample strength. The DEM results are in very good agreement with the experimental results on the development of both shear stress and volumetric change. Further, lateral normal stresses in the simple shear in the DEM study can be obtained, enabling the computation of principal stresses under two dimensional shears. Fabric developments of samples can also be studied in the DEM simulation.

Dielectric properties of soils are of paramount significant in geo-environmental applications. Despite the well-studied effects of water content on the soil dielectric permittivity, the simultaneous effects of wet contamination and dry density have remained relatively untapped. This paper investigates these effects on three soils (silica sand, kaolinite and bentonite) with distinct specific surface area at 1 GHz. The results of the experimental program were evaluated against the Birchak model. It was found that the relationship between dry density and dielectric properties is directly proportional for soil/oil mixtures. Furthermore, the geometrical parameter in the Birchak model was found to be relatively constant for soil/oil mixtures, presumably due to the non-polar molecular structure of the contaminant under study (i.e., oil). For soil-water mixtures, however, it is suggested that this parameter is a non-constant value and is a function of dry density.

In the case of a sudden change in the geology in front of the Tunnel Boring Machine (TBM) during mechanized tunneling processes, non-appropriate investigation and process adaptation may result in non-desirable situations that can induce construction and machine defects. Therefore, subsurface anomalies detection is necessary to trigger alarm to update the process. This paper presents an approach for geological anomaly detection using data produced by the TBM. The data observations are continuously produced at a motion of 10 to 15 s from hundreds of sensors around the TBM. Unsupervised machine learning techniques are applied to analyze the online streaming data. As a result, a model, which is able to learn the system characteristics from normal operational condition and to flag any unanticipated or unexpected behavior, is established. The proposed approach has been tested on the data of the Wehrhahn-Linie metro project in Düsseldorf in Germany. The model can accurately detect the presence of concrete walls in the ground domain with a distance up to around one meter before the TBM approaches the walls. The developed method can thus be used as a monitoring system for ground risks detection to ensure safe and sustainable constructions in mechanized tunneling.

The Seismic Safety and Resilience of Schools in Nepal (SAFER) project has an important aim of producing improved tools for geotechnical and earthquake engineers to assess seismic hazard in the Kathmandu Valley, Nepal. Geo-databases have the potential to offer geotechnical practitioners means to improve a-priori predictions of important soil parameters in geotechnical design. In this paper, some recent work to develop a new database of geotechnical information (SAFER/GEO-591) including shear wave velocity measurements is reported. Attempts to develop new transformation models to better predict shear wave velocity from more basic parameters such as SPT-N values are presented. Use of kriging to better map shear wave velocity for the study area is recommended as a suitable alternative to the presented correlations.

Soils and rocks are natural materials, and they are affected by many spatially varying factors during their complex geological formation process. Geotechnical property therefore exhibits spatial variability, which is site-specific. The site-specific spatial variability of geotechnical property is often non-stationary (e.g., with spatial trend) and may not follow a Gaussian distribution. On the other hands, investigation data from a site is often sparse and limited in geotechnical engineering practice due to time and budget constraints. This leads to a great challenge of how to properly quantify the non-stationary and non-Gaussian geotechnical spatial variability in a specific site from sparse measurements. A novel method is presented in this paper to address this challenge. The method is based on Bayesian compressive sampling and Karhunen-Loève expansion, without assumptions of parametric trend function form, parametric auto-correlation function form and marginal probability distribution type.

In site characterization and modelling, subsurface spatial variability is often characterized using the scale of fluctuation, Θ, in the horizontal and vertical directions. Typically, these scales are estimated by statistically fitting an appropriate autocorrelation function to the CPT data from the layer of interest. These statistical techniques are data intensive requiring significant amounts of data to provide accurate estimates. While in the vertical direction, along the CPT, the available data is abundant, the amounts of data in the horizontal plane is limited to the number of CPTs undertaken; therefore, these traditional approaches can adequately estimate vertical scales, Θv, while the horizontal estimate, Θh, is difficult to obtain.This paper aims to expand on the previous work of the Author, in using a neural network-based approach to estimate these spatial statistics from CPT data. This study expands the work from simple 2D domains, modelled as random fields, to 3D, and more realistic site surveys, and the methods are compared.

Emergence of non-uniform deformation modes or instabilities often encountered in laboratory “single element” tests have been examined in this study across various “length-scales” of granular media viz, continuum, discontinuum and laboratory biaxial “element” test level. With inhomogeneity outset, the mechanical response no longer remains a true representation of the material behaviour. It rather portrays the system response with the boundary conditions and the evolving instabilities. Instability onset within a transient undrained continuum elastoplastic framework is found to be a mesh-dependent phenomenon. This “pathological mesh-dependency” of classical Cauchy continua is addressed with the aid of Level Set Discrete Element Modelling of Flexible Boundary (FB) Plane-Strain (PS) tests that takes into account the actual grain morphology into consideration. The micromechanical observations are found to be in good qualitative agreement with the macromechanical “element” response of FB-PS tests. Interestingly, the microstructural arrangement or the grain fabric acts as the triggering mechanism behind the shear strain aggregation in localized zones within the sand specimen. Alternately, the evolution of heterogeneities may be considered as “disturbances” diffused within the sand specimen that later manifest into localized zones of shear strain accumulation on gradual shearing.

The intergranular strain concept allows a consideration of the deformation history of the soil in constitutive models. In this contribution this concept is applied on the constitutive model barodesy. This combination allows the usage of the asymptotic state boundary surface of barodesy, which helps to reduce the overshooting effects of the intergranular strain model. The finite element implementation of the model is presented, including a strength reduction procedure to calculate the overall factor of safety. The functionality of the model is verified in the simulation of the excavation of a construction pit. The influence of the intergranular strain concept on resulting deformations as well as on the calculated factor of safety is investigated.

This paper presents a study on numerical simulations of clayey soil using nonlinear elastic viscoplastic (EVP) model. The nonlinear EVP model is based on Yin (1999)’s nonlinear creep equation in which a creep strain limit is considered. The model in the general stress-strain condition is implemented by finite element program Plaxis and is used to simulate clayey soil under Berthierville embankment. Good agreement between simulated and measured settlements and excess pore pressure validates the feasibility of the nonlinear EVP model. Furthermore, a parametric study is presented to demonstrate the importance of creep strain limit.

The paper focuses on the numerical prediction of the stress-strain behaviour under triaxial compression of a sandy clay soil improved with an environmentally friendly binder at the short-term (28 days) and long-term (90 days) using a kinematic hardening constitutive model. The binder was synthesised by ground granulated blast furnace slag (GGBS), an industrial by-product (IBP) of the steel industry, and sodium hydroxide (NaOH). The model, which is being used to reproduce artificially cemented soil behaviour for the first time, was able to successfully capture the smooth elastic-plastic transition response in the unimproved soil, while a ductile response observed in the improved soil prior to a fragile post-peak and residual states was also well predicted by the model.

Based on the effective stress concept of continuum damage mechanics (CDM) and the assumption of a small deformation, a constitutive model of salt rock coupling viscoelasticity, viscoplasticity, and viscodamage is developed based on the Kelvin-Voigt model, Duvaut-Lions model and viscodamage model. The constitutive model is calibrated by experimental data of salt rock to determine the viscoelastic, viscoplastic, and viscodamage model parameters. The verification indicated that this constitutive model can reasonably model the primary creep, steady-state creep, accelerated creep in constant stress loading. Based on user material subroutine UMAT in Abaqus software, a three-dimensional finite element implementation of this constitutive model is realized. The constitutive model can effectively simulate mechanical responses including the long-term time-dependent deformation and damage evolution of geological bodies of salt rock and successfully conduct stability evaluation and failure analysis of cavern of large time span. This will provide a safety guarantee for underground energy storage.

The macroelement approach, commonly employed to study soil-structure interaction problems, stems from the intent of describing the mechanical response of a complex system by means of a single upscaled constitutive relationship, relating generalized stresses and strains.In this paper, the authors introduce a new generalized constitutive relationship capable of reproducing the undrained mechanical response of both foundations and tunnel cavities. To this aim, the authors performed a series of numerical analyses. The numerical analyses results are interpreted to put in evidence that the mechanical response of both systems is governed by a “structural hardening” associated with the spatial propagation of the yielded soil domain that is influenced by the relative position of the structure with respect ground surface: when the yielded soil domain reaches the ground surface a failure condition is got. To take into account this aspect, the authors introduced in the constitutive relationship two independent plastic mechanisms, the former one defines the spatial propagation of the yielded soil domain (structural hardening), whereas the latter one the failure mechanism due to the boundary condition. The comparison of the constitutive relationship predictions with the finite element numerical data demonstrates that the proposed model is capable at the same time of satisfactorily reproducing the mechanical response of shallow/deep tunnels and shallow/embedded foundations.

Many Offshore wind projects are being developed nowadays. Design of monopiles, the most used foundation system for the offshore wind turbines, requires a special attention in the characterization of the soil-pile interaction. The main feature induced by axial cyclic loading at the level of soil-pile interface is the decrease in normal stress. Hence, evolution of soil-pile interface normal stress may affect shearing mechanism of the pile. Using a finite difference software, this paper presents an alternative method to simulate the interface response. Curves illustrating the evolution of normal displacement in function of shear displacement, extracted from experimental shear tests, were directly injected as disturbed relative normal displacement in the interface nodes. Therefore, shear stresses of the interface nodes are directly corrected by equilibrium procedure, based on the principle of elasto-plasticity. Calculation results present good agreement with experimental results.

The Intergranular Strain Anisotropy (ISA) model by Fuentes and Triantafyllidis (2015) corresponds to an extension for the conventional hypoplastic models to account for the effects observed under cyclic loading. Similar to the Intergranular Strain (IS) theory proposed by Niemunis and Herle (1997), this extension enhance the hypoplastic models in many aspects as increasing the stiffness and reducing the plastic strain rate on cyclic loading. The present work is devoted to extend and evaluate a constitutive model for anisotropic clays under cyclic loading. The reference model corresponds to the anisotropic hypoplasticity for clays by Mašín (2014), which accounts for an elastic tensor depending on the bedding plane’s orientation. The proposed model results from extending the hypoplastic model with ISA. For validation purposes, different simulations under monotonic and cyclic loading were performed with an anisotropic kaolin clay. It was found that the proposed model accurately predict the experimental results under a wide range of strain amplitudes.

The numerical implementation of a recently developed thermomechanical constitutive model for fine-grained soils based on hyperelasticity-hyperplasticity theory (Golchin et al. 2020), is presented. A new unconventional implicit stress return mapping algorithm, compatible with elasticity derived from Gibbs (complementary) energy potential, in strain invariant space, is designed and the consistent tangent operator for use in boundary value problems (such as in the finite element method) is derived. It is shown that the rate of convergence of the stress integration algorithm is quadratic. The numerical results are in good agreement with available data from thermomechanical element tests found in literature.

A numerical model based on two-dimensional Discrete Element Method (DEM) has been used to study the compression of rock under different loading rates, and a visco-elastic Flat-Joint constitutive law has been developed to describe the dynamic behavior of bonded contact between particles. The improved DEM model consists of two parts: (1) Force-displacement equations are derived on the basis of a visco-elastic constitutive model which is formed with a spring and a Maxwell system, and (2) A dynamic fracture criterion of bond is develop to define the increase of strength. The controlling variable method is used to analyze the impact of visco-elastic parameters on the dynamic mechanical properties. With different parameters this model could be used to simulate the compression of rock materials under different loading rates. As an example, visco-elastic parameters of granite are identified, and the obtained stress-strain curves well reproduce the laboratory results.

In the present study, we propose a new elastoplastic constitutive equation with consideration of the hydrate morphology. Then, using the proposed constitutive equation, we investigate how large the strength and the volumetric strain change by the change in the hydrate morphology with the fixed total hydrate saturation. The result indicates that the strength and the positive dilatancy becomes the larger in the case where the morphology of the cementing-type is more dominant. Moreover, the strength curves predicted by the proposed model is in good agreement with the past experimental research. This makes it possible to predict the strength and the deformation behavior of MH-bearing sediments even in the area where there is a lack of research data.

Cyclic laboratory tests on sand are generally performed either under drained or undrained conditions, while the actual in-situ conditions typically are partially drained. It is therefore necessary to make some assumptions in order to use the cyclic properties obtained from these tests in design. To check the consequence of these assumptions some special cyclic triaxial tests with different degree of drainage are therefore carried out at NGI. This paper presents the results from these tests together with a detailed interpretation. It is shown that the measured accumulated pore pressure and volumetric strain may be very well back-calculated by three basic equations. The accumulated volumetric strain during a cycle is the key material property, together with a represented bulk modulus and the flow resistance of the special filter device. Based on the results, it is concluded that more study is required to fully understand the governing properties that controls the accumulated strains, especially the effect of previous stress or strain history.

An evaluation of the slope stability in FEM software is often performed as a complementary analysis to the plastic deformation analysis. The slope stability represented by the factor of safety is, in the most of FEM software, evaluated by the reduction of the strength parameters $$c$$ c and $$\varphi $$ φ of the Mohr-Coulomb model. Consequently, the lack of the strength reduction method for the advanced constitutive models inevitably leads to its substitution by the Mohr-Coulomb model and necessity of a further recalibration of the constitutive model parameters. Therefore, new method of the strength reduction was developed for the Masin’s hypoplastic model for clays. A subsequent return of a deviatoric stress to the asymptotic state boundary surface is implemented in this method so that all the integration points within the FEM remain within asymptotic state boundary surface. Predictions of the newly developed method is thoroughly testes from the quantitative as well as qualitative point of view and eventually compared with the predictions of the Mohr-Coulomb model.

Granular rock salt as a strong time dependent and stress dependent material behavior. For the numerical simulation of this material behavior a new viscoplastic constitutive law was developed and implemented into a Finite-Element-(FE-)Program. The numerical simulation of the material behavior is necessary for the analysis of the stability and the serviceability of large tailings heaps and infrastructure construction in the area around these heaps. The tailings heaps occur during the production of potassium fertilizer and consist of granular rock salt. The paper explains the mathematical formulation of the new viscoplastic constitutive law which is called ViscoSalt 2017 and the implementation into a FE-Program. The salt mechanical material parameters for this constitutive law are determined by load controlled triaxial creep tests and strain rate controlled triaxial fracture tests. For the verification of the new constitutive law and the correct implementation into the FE-Program the back analysis of the triaxial tests is used. Now the new constitutive law is used for the analysis of the stability and the serviceability in engineering practice.

Coarse granular materials such as railway ballast used in the shallow layers of railway tracks are often subjected to moving loads, which cause complex stress conditions involving the rotation of principal stress axes. Further, these materials are involved with particle breakage, which is also affected by the stress changes in the track. Computer modelling incorporating constitutive relationships is an effective technique for assessing the deformation behaviour of coarse granular materials under such conditions. This paper presents a constitutive model for coarse granular materials in multilaminate framework, considering classical plasticity and critical state concepts. A criterion for particle breakage in multi-laminate framework and its effect on the stress-strain behaviour has been incorporated using constitutive relationships in multilaminate framework. The influence of minor principal stress and principal stress rotation on the stress-strain, volumetric strain and particle breakage of these materials have been discussed.

The history of geological deposits is often characterized by complex processes which strongly modify the mechanical behavior of rock masses. The intrinsic orientation-dependent properties connected with the presence of bedding planes induces anisotropic characteristics in the rock masses. In order to model their mechanical behavior, a 3D Hoek & Brown (HB) yield criterion has been used to introduce material symmetries through cross anisotropic elasticity and by means of the Uniaxial Compression Strength (UCS) expressed as a function of the bedding plane inclination. The material degradation and the resulting brittle mechanisms are simulated with a softening rule aimed to introduce a decreasing evolution of the Hoek & Brown parameters, thus enabling to model the post-peak behavior of the intact rock, as well as its non-linear dilation trend. This constitutive framework has been employed to evaluate the influence of preferential directions on the failure process by solving plane strain compression tests in PLAXIS 2D in which a viscous regularization technique is implemented to avoid the pathological mesh-sensitivity of the numerical solution during the formation and propagation of shear bands.

The authors focus on constitutive modelling of solid to fluid phase transition in granular soils. A new constitutive approach conceived to capture such a transition in either dry or saturated granular material is validated against DEM numerical simulations for steady conditions and discussed in the framework of the μ-e-I rheology. The approach considers granular and liquid phases separately, assuming the two phases work in parallel, as it is according to the effective stress principle in case of quasi-static conditions. The authors emphasize the importance of considering separately the two phase contributions and the limitations of the μ-e-I rheology, in particular in relation to the volumetric response of the material.

In geotechnical engineering, boundary value problems are often idealized under plane strain conditions. It is therefore of particular interest to investigate constitutive models with respect to plane strain failure. In this contribution, previous studies on elastoplastiticity are linked to the here presented studies on hypoplasticity and barodesy. The parameters of the models are chosen in such way that strength predictions coincide for drained axisymmetric triaxial conditions. The plane strain strength predictions differ due to different failure surfaces and due to different deviatoric directions of stress paths under plane strain conditions. The influence of the initial stress state on peak strength is discussed. As barodesy and hypoplasticity are formulated within the asymptotic state concept, critical stress states are independent of the initial stress state. The results of element test observations are related to a strength reduction analysis.

Identification of the correct locations of the current stress and stress reversal points in the 3D stress space is crucial for the successful application of a 3D bounding surface model under complex multiaxial loading conditions. Any miscalculation of the locations of these points can markedly alter the location of the image point on the bounding surface, and consequently affect the model predictions. The main objective of this paper is to present a generalised mapping rule for the identification of image point in 3D bounding surface plasticity models. The mapping rule is based on determination of the exact locations of the stress points in the principal stress space using the eigenvectors of the stress tensors. The mapping rule is implemented in a 3D bounding surface plasticity model to capture the behaviour of soils under multiaxial loading conditions. The simulation results are presented and compared with existing experimental data to validate the proposed model.

In this paper, an advanced critical state compatible Bounding Surface (BS) plasticity model for the cyclic response of clays is developed based on a prior version of a Simple Anisotropic CLAY plasticity (SANICLAY) model. With the proposed model, named SANICLAY-BS and abbreviated as S-BS, it is possible to capture the presence of a cyclic stress threshold above which large strains develop leading to the effective stress failure. In addition, the number of cycles to failure can be controlled for a wide range of applied CSRs. Peculiarity of the formulation is the incorporation of a novel activation mechanism into the destructuration law. Model performances in capturing a cyclic resistance curve are shown against the experimental data of Cloverdale clay.

Argillaceous rocks have a complex and heterogeneous structure at different scales. At the scale of the mineral inclusions embedded in a clay matrix, the deformation generally induces microcracking and material damage. Modelling the latter requires to take into account microscale characteristics and their effect on the micromechanical response. This response can be used in double scale approach to predict material behaviour at larger scale. Thus, heterogeneous microstructures of a claystone are generated with a distribution of morphological properties satisfying experimental observations. The overall microscale material behaviour under solicitation is obtained by numerical homogenisation. Then, the variability of the material response is studied with regard to small-scale characteristics. In terms of deformation and failure, a dominant shear deformation mode and decohesion between grains are observed. The decohesion induces microcracking in the microstructure and strain softening of its overall response.

The nonlinear behaviour of unsaturated soils is governed by the fully coupled hydro-mechanical phenomenon due to the irrecoverable movement of particles and fluids. It is usually accounted for in constitutive modeling using separate evolution rules for plastic deformation and saturation, linked with two yield conditions for stress and suction. In this paper, a new generic thermodynamics-based approach is developed to provide a more rigorous way to capture these underlying mechanisms. A special form of dissipation potential leading to strong inter-dependence of mechanical and hydraulic internal variables is used for the derivation of a single yield surface. A specific critical state model using a small number of identifiable parameters is derived from the proposed formulation. Its capabilities in predicting the drained and undrained experimental results are investigated to highlight the applicability of our approach.

The expansion and penetration of the structural planes under stress is an important cause of rock mass failure. Based on the principle of linear elastic fracture mechanics, a multi-crack propagation simulation algorithm of higher-order numerical manifold method (NMM) is proposed. The singularity of the crack tip displacement field is considered by adding key terms of the crack tip displacement field function to the basis function of the NMM. The stress intensity factor at the crack tip was calculated by J integral. The cracking and propagation directions of I-II mixed cracks are judged by the maximum circumferential tensile stress criterion. Hypothesis-modified multi-crack propagation algorithm is used to solve the problem of multi-crack propagation. When a single physical cover contains two or more crack tips, the influence of the singularity of the displacement field at each crack tip will be considered simultaneously in the physical cover displacement function. For those integral functions which do not conform to the simplex integral form, Taylor series expansion method is used to calculate the approximate solution. The rationality and accuracy of the calculation method are verified by numerical simulation of two classical static crack propagation problems and physical model test.

The numerical simulations of granular materials, in the framework of continuum mechanics, is quite challenging since the constitutive model should be capable of reproducing the transition from solid- to fluid like regimes and vice versa.In this paper a constitutive model, valid under general three dimensional evolving conditions, and capable of describing the material response under both quasi-static and dynamic regimes is presented. The model is calibrated by employing a series of true triaxial DEM numerical simulations performed on a periodic cell. The comparison between model predictions and DEM results highlights that the capability of the constitutive relationship of taking into account the dependence of the mechanical behaviour of the dry granular material on Lode angle, strain rate, void ratio and confining pressure.

In the present study the implications of a thermodynamic-based constitutive framework on the mechanical behaviour of soils are critically analysed. The primary advantage of this approach as compared to the traditional hardening plasticity is that the models are guaranteed to obey the laws of thermodynamics. Furthermore, the use of potential functions allows to introduce some crucial ingredients of the mechanical behaviour of soils that directly affect the shape of the yield surface and the flow rule of the model. To illustrate the above features, different forms of elasto-plastic coupling are presented and their implications on the response of the models are explored with reference to a series of numerical simulations.

In many geotechnical problems, failure conditions involve the formation of a narrow shear band where shear strains localize. In this study, the thickness of the shear bands was indirectly determined based on the results of constant normal load direct shear (CNL) tests carried out on sand specimens reconstituted at three different relative densities. The adopted procedure also allows to evaluate the soil deformations within the shear band and, thus, to correctly locate the critical state line in the compressibility plane. The results of the CNL tests were used to calibrate the Severn-Trent model, an advanced constitutive model proposed by Gajo and Wood, apt to well-reproduce the mechanical behaviuor of sands from small to large strain levels. The predictive capabilities of the model were confirmed by the good agreement with the experimental data obtained performing constant normal stiffness direct shear (CNS) tests. Finally, the shaft bearing capacity of a bored pile embedded in a homogeneous soil layer was numerically evaluated and compared to the one predicted using a less sophisticated (classical) approach.

When the discrete element method (DEM) is used to simulate rock dynamic experiments with split Hopkinson pressure bar (SHPB), it is difficult to obtain the result that the numerical and experimental rock specimen will have the same rapid growth of dynamic compressive strength when strain rate increases. To solve this problem, an improved contact bond model of DEM considering the crack propagation velocity was proposed. Then this new contact bond model was used to simulate the rock dynamic experiments with SHPB under different loading rate. Compared with the simulation results of the original contact bond model, the simulation results of this new contact bond model are closer to the experimental results.

Hydrostatic and one-dimensional compression behavior of granular material are closely related to particle breakage. Tangent constrained modulus may decrease with loading because of the significant crushing of particles. However few compression models could represent this behavior. The traditional semilog linear compression model only shows increasing modulus. In this article, an elasto-plastic compression model is proposed. The plastic strain is divided into plastic hardening strain and plastic breakage strain, the former is due to particle packing, interparticle slip, and rotation, while the latter is due to breakage of asperities and whole-particle fracture. Plastic breakage strain is linked with particle characteristic strength following a power-law particle size effect. The model is verified by experimental results from the literature.

The contribution is motivated by the need for numerical analysis of flow in fractured porous media, i.e. rocks in the geo-engineering applications. We describe development and testing of numerical methods for simulation of (coupled) flow and deformation in a fractured porous environment. The hydraulic behavior is described by the Darcy flow in the porous matrix and the Poiseuille flow in the fractures. The fractures are considered as domains of reduced dimension. Both the matrix and fracture flow are interconnected by the flux through the fracture walls. The mechanical behavior is described by the linear elastic deformation of the porous matrix with contact conditions on fractures. In this way, it allows fracture opening and closing with the constraint on non-penetration. The slip effects are not considered. We consider both steady-state and time-dependent problems.

A multiscale model is developed for the modelling of coalbed methane (CBM) production. CBM recovery is known to be a highly coupled and multiphase problem. The finite element square method is used to integrate a fracture-scale model in a multiscale scheme. This method consists to localize the macroscale deformation to the microscale, then resolve the boundary value problem on the microscale with finite elements, then homogenize the microscale stresses to compute macroscopic quantities, and finally resolve the boundary value problem on the macroscale with finite elements. This approach has the advantage that it does not require to write some constitutive laws at the macroscale but only at the REV-scale. The multiscale model is therefore appropriate for reservoir modelling. The model is developed and implemented in a finite element code and the simulation of a synthetic reservoir is considered.

The effect of temperature on the mechanical behaviour of geomaterials is relevant in a number of geotechnical applications including low enthalpy energy geostructures. Mechanical response of clays upon heating is not always intuitive, and this includes the volumetric ‘collapse’ observed in normally consolidates clays. This paper reviews the effect of temperature on clay particle interaction forces (Colombian, van der Waals and mechanical) in the attempt to elucidate the micro-mechanisms underlying temperature-induced behaviour observed at the macroscale.

Understanding the soil in the unsaturated state means, among other things, understanding how the presence of water transiently affects the physical and hydraulic properties of the soil environment. This is extremely important in solving soil resistance problems involving water, such as the stability of land slopes. Among the hydraulic properties that are essential in understanding water flow in porous media, flow velocity and soil water retention capacity are the most important defining characteristics of the hydraulic behavior of the medium. In the physical model used in this study, these characteristics were represented by the advective velocity and hydraulic diffusivity parameters that were obtained from the infiltration data by conducting tests on the latosol columns. The main results obtained can be used for the estimation of the wetting front advance given the uncertainties of the variables obtained from the confidence envelope curves.

This study presents a 3-D particle simulation model based on the discrete element method for the weathering processes of geomaterials. We modeled their properties of swelling and shrinking by changing the diameter of spherical region which is consisting of clay minerals and several sand particles. A series of simulations of deformation of mudstone under wet-dry cyclic condition was performed so that the applicability of the proposed model can be validated.

The founding principles of consolidation of soils was set by Terzaghi dating back to 1923. This theory focused on the interplay between pore pressure dissipation and effective stress increase on fully saturated soils, and works by means of the diffusion theory in one-dimensional flow conditions under constant static external loading. The theory remains to use still and works as the funding stone for all subsequent formulations in the international theory. It is noted, that all consolidation formulations reduce to the original Terzaghi’s theory (as they comprise its extensions) for one-dimensional flow under constant static loading on fully saturated conditions. This paper revisits the consolidation theory and derives the governing equations via a rigorous integration of mass conservation, which accounts for instantaneous void ratio alterations within the continuity equation by including an additional term in the consolidation coefficient. This gives shorter consolidation times compared to Terzaghi’s original expression. The derived equations reduce to Terzaghi’s with appropriate manipulation. The paper compares the 1D consolidation coefficients to Terzaghi’s and other formulations, and proposes a simplified solution of the consolidation equation based on regression, accompanied with a workaround to include the rigorous approach in commercial (FEM, FDM) codes within the consolidation governing equations.

Swelling clays are found all over the world, and the damage caused to the infrastructure in swelling clay areas is estimated to be about 20 billion dollars annually in the United States. Compressibility and shear strength are critical properties of soils that are necessary for the design of infrastructure. Our group has shown that molecular interactions between clay and fluids have a dramatic impact on the macroscale properties of swelling clays. The permeability of the clay increases about 500,000 times when the permeating fluid in the Na-montmorillonite clay is changed from polar fluid water to low polar fluid. This change results from clay–fluid molecular interactions as well as the differences to the microstructure caused by these interactions. In the current work, the compression and shear strength of the clay interlayer is studied by changing the polarity of the fluid in the interlayer and applying compressive and shear stresses using steered molecular dynamics. The results show the strong influence of normal stress as well as fluid polarity on compressibility and shear response of the clay at the molecular scale. The results demonstrate that clay–fluid molecular interactions play a crucial role in the macroscopic compressibility and shear strength of swelling clays.

Pore fluid composition strongly influences the mechanical behavior of clays, impacting both on their volumetric and shear response. Accounting for this aspect is crucial for engineering applications where changes of the chemical composition of the pore fluid are anticipated, such as transport through engineered barriers for the containment of pollutants, or slope stability of natural formations rich of clay minerals subjected to freshwater infiltration. In this work, a chemo-mechanical model capable of reproducing the response of medium to low activity clays under both mechanical and chemical loading paths is presented. The model is developed starting from the interpretation of experimental evidences in an elastic-plastic framework. Chemo-mechanical coupling is introduced both in terms of stress variables and hardening law. In particular, the formulation is specialized to variations of salt concentration, introducing osmotic suction as a chemical stress variable. The model was implemented in a constitutive driver for the integration at the REV level of the incremental constitutive equations, thus allowing for its validation against literature data.

In this study, an energy based hydro-mechanical model and computational algorithm for the problem of hydraulically driven fracture networks developing in naturally fractured impermeable media is developed. The model is based on non-differentiable energy minimization for the dynamic deformation and fracture of the body coupled with mass balance of fluid flow within the hydro-fractures. Time-discontinuity induce spurious crack opening velocity fields which lead to nonphysical solutions for the coupled fluid pressure field defined locally along the crack faces. The use of a time-continuous fracture model, such as the present non-differentiable energy minimization approach, is crucial for the numerical soundness and stability of the hydraulic fracture propagation algorithm. A discontinuous Galerkin finite element formulation is implemented, in which every element edge in the mesh is a potential site of hydro-fracture initiation and propagation. The presence of pre-existing natural fractures, as a common challenge in nearly all geological formations, are modelled with desirable flexibility by simply assigning different fracture properties to the element edges defining the natural fractures. Using the graph theory principles, a search algorithm is proposed to identify, among all, the sub-set of cracked interfaces that form the interconnected hydraulically loaded fracture network. Robustness of the proposed computational algorithm and its versatility in the study of hydraulic fracturing are shown through several numerical simulations.

In France, a deep geological disposal for high-level radioactive waste (HLW) and intermediate level long life radioactive waste (IL-LLW) called Cigéo is planned to be constructed in a deep Callovo-Oxfordian claystone (COx) formation, if it is authorized. The heat emitted from the waste leads to a temperature increase in the low permeability host porous formation, which induces a pore pressure increase essentially due to the difference between the thermal expansion of the pore water and the solid skeleton. This study aims at assessing the effect of nonlinear behavior of COx, especially creep, on the THM response of the HLW repository. Different approaches, from perfectly Mohr-Coulomb model and simplified time-dependent models (power and Norton) to an advanced model, have been considered. These approaches lead to common conclusions that (1) the creep reduces the pore pressure increase compared to poro-elastic approach; (2) the nonlinear instantaneous response in the near-field resulting from cell drilling does not affect the THM response in the far field and (3) the thermal load does not induce any supplementary damage of the host rock.

Peridynamics (PD) has received more and more attention in dealing with discontinuous problems. In this theory, the crack initiates and propagates arbitrarily without any extra remeshing strategy or propagation criteria. To improve the deficiency of computationally more demanding, PD is coupled with the classical finite element method (FEM). One symmetric double notched tension test is taken to verify the effectiveness of PD as well as the coupling method. The influence of parameters on the damage evolution is investigated based on this tension test. It is found that PD is capable to describe the crack propagation process in elastic-brittle materials. The overall force-displacement response is influenced by the size of loading increment. Importantly, the structural response is sensitive to the horizon radius. The results are approximately convergent when the horizon radius is no less than three times the specific mesh size. For a particular (three times) ratio, the peak force and the peak displacement are negatively affected by the mesh size. The critical energy release rate positively influences the peak force as well as the peak displacement. The Young’s modulus has a positive effect on the peak force but the negative effect on the peak displacement.

An integrated model of fully coupled two-phase surface and subsurface flow was built by using GETFLOWS (GEneral purpose Terrestrial fluid-FLOW Simulator), a finite difference fluid-flow numerical simulator, to perform quantitative evaluation of exchange of overland water and groundwater, subsurface seepage and surface water runoff on the hillside slope surface. Surface water on hillside slope was modeled as open-channel flows and coupled with air-water two phase Darcy flow for underground permeation by using a generalized flow formula. In this study, a simple verification model was firstly built and validated by the extensively used Abdul and Gillham example with simulating the coupled surface and subsurface flow. Then a three-dimensional integrated model of surface and subsurface flow was built by using the DEM (Digital Elevation Model) of a natural mountain slope in Hokkaido, Japan, where runoff induced several debris flows and slope failures occurred during Typhoon 10 hit Hokkaido in 2016. Finally the results of integrated simulation was compared with the results from two-dimensional (2D) impermeable plane flow simulation using Nays2D Flood solver of the iRIC software in the same region. The results suggest that on the hillside slope upstream of embankment, the infiltration rate is equals to the rainfall intensity at the beginning of the rainfall event. After the rainfall intensity is greater than the soil infiltration capacity, runoff is generated. The runoff from upstream allows more water to infiltrate into the embankment causing the possibility of slope failure at the exit of the valley to be much greater than other locations along the highway.

In this paper a relationship between elastic anisotropy, as typically observed in clayey soils subjected to shear wave propagation tests, and plastic anisotropy, detected at yielding and leading to rotated yield loci, is proposed. Furthermore, elastic and plastic anisotropies exhibit an evolving character, as a consequence of the evolution of the fabric of the material induced, for example, by irreversible straining. First, the theoretical implications of the above evolving character of fabric, which leads to a new form of anisotropic elasto-plastic coupling, are investigated. Then, a strategy is proposed to take advantage of such a coupling to more effectively initialise the internal variables of any non-isotropic hardening plasticity model. This latter aspect is of crucial importance when numerically analysing the response of a whole deposit of soil, as for each sub-stratum it is mandatory to identify the initial orientation of the yield locus.

The occurrence of a sinkhole in an area may compromise the safety of the existing structures and infrastructures. Therefore preventive solutions are necessary. Recently, a coupled DEM-FEM numerical model has been developed to better account for the failure mode of reinforced soil layers during cavity openings and the interaction between the collapsed soil and the geosynthetic sheet. An experimental campaign on a small scale trapdoor model gave the possibility to validate the numerical model. In that case, a usual linear elasto - perfectly plastic Mohr Coulomb constitutive law and its failure criterion have been chosen to represent the cohesive soil. A good agreement with the experimental results has been observed. To be able to reproduce the behavior of various cohesive materials, an advanced constitutive law has been tested in place of the usual model based on the Mohr Coulomb criterion.

Increased traffic and environmental loads necessitate re-evaluation of the stability of existing road and railway embankments built on soft sensitive clays. Thus, the current mobilised undrained shear strength needs to be quantified. A methodology to evaluate changes in undrained shear strength under embankment loading is developed and applied for the case of Haarajoki test embankment. The methodology combines boundary value modelling of embankment loading with integration point level stress path probing using the Creep-SCLAY1S model. The changes in the stress state and the relevant state parameters resulting from the boundary value modelling enable the quantification of the mobilised undrained shear strength. The results indicate an increase in the undrained shear strength up to 17% in the most affected clay layer. The high pre-overburden pressure in the top of clay deposit prevents significant changes in the undrained shear strength in the case of Haarajoki. Thus, when assessing changes in the undrained shear strength, one of main parameters to determine is the initial preconsolidation pressure.

Natural frequency wandering is a phenomenon observed in dynamic structural behavior. Temporary and permanent changes in the determined natural frequencies have been shown to be related to such factors as weather conditions, structural damage or soil-structure interaction. Furthermore, some studies indicate temporary natural frequency wandering during seismic events. This temporary wandering and recovery of the initial pre-event natural frequency has not been specifically explained yet. Some studies indicate soil-structure interaction as a potential reason causing the observed structural behavior. This paper presents a numerical finite element study and aims at investigating the influence of soil-structure interaction on the observed natural frequency wandering induced by seismic actions. The advanced soil constitutive model Severn-Trent is used and an example structure built on saturated granular soil is analyzed. A temporary drop in the natural frequency is shown soon after the earthquake followed by regain to the initial value with time. The results show that generated pore pressure, its dissipation over time, reduced mean effective pressure and nonlinear stress-strain behavior are linked with temporary changes in natural frequencies.

Constitutive modeling of granular materials such as sands, non-plastic silts, and gravels has been significantly advanced in the past three decades. Several new constitutive models have been proposed and calibrated to simulate the results of various laboratory element tests. Due to this progress and owing to the surge of interest in geotechnical engineering community to use well-documented constitutive models in major geotechnical projects, a more thorough evaluation of these models is necessary. Performance of the current models should be particularly evaluated in the simulation of boundary value problems where stress/strain paths are much more complex than the element tests performed in laboratory. Such validation efforts will be an important step towards the use of these models in practice. This paper presents the results of an extensive validation study aimed at assessing the capabilities and limitations of a two-surface plasticity model for sands in two selected boundary value problems, i.e. lateral spreading of mildly sloping liquefiable grounds. The results of a large number of centrifuge tests conducted during the course of four consecutive international projects known as Liquefaction Experiments and Analysis Project (LEAP) are used in this validation study. The capabilities and limitations of the two-surface plasticity model, initially calibrated against element tests, will be carefully assessed by comparing the numerical simulations with the results of the centrifuge tests from recent LEAP projects.

For the design of structures with shallow foundations in the vicinity of active faults sophisticated numerical analyses are a prerequisite. Current practice employs isotropic constitutive models for the simulation of the foundation soil layer. For enhanced accuracy, these models may also include strain softening, as well as a non-associated flow rule if this soil is granular, but the anisotropic nature of soils is not considered. In order to study how much sand fabric anisotropy affects the fault rupture-foundation interaction problem, this paper employs the finite difference method and 3 constitutive models of various complexity, including a new fabric-based model for sand named SANISAND-FR. All models are calibrated to give identical results under triaxial compression. The analyses show that the assumption of isotropy affects significantly the fault rupture–foundation interaction, and hence the expected response of the structure (e.g. rigid body rotation) due to rupture surfacing. This means that the design of structures against fault rupturing is of doubtful accuracy if models are used that do not consider sand fabric anisotropy.

Conventional design methods estimate the bearing capacity of shallow foundations by considering the soil as an isotropic elasto-plastic medium with an associated flow rule. State-of-the-art numerical analyses of such foundations employ constitutive models that consider strain softening and a non-associated flow rule if the soil is granular. Such analyses lead to increased accuracy over the conventional design methods, i.e. relatively lower bearing capacity, yet they do not consider the effect of soil fabric anisotropy. In order to investigate this effect, analyses of the bearing capacity of shallow foundations are carried out, by means of the finite difference method using the SANISAND-FR constitutive model that is fabric-based and verified against element and centrifuge test results. A comparison is made here between the analyses with this model and its variant that does not consider fabric effects, after their calibration to give identical results in the usually available triaxial compression tests. The comparison shows that neglecting sand fabric anisotropy, along with model calibration on triaxial compression tests, lead to an overestimation of the bearing capacity of footings, i.e. to un-conservative design.

To model the elasto-viscoplastic soil response, even in one-dimensional problems, a significant amount of parameters are required. These parameters interact, making it difficult to fit the parameters directly. A model is fit to data by numerical optimization with a number of local and global algorithms. The evaluation showcases that the correct parameters can be found efficiently by optimization, but that there is a significant risk of converging to a local minimum far removed from the correct solution as the dimensionality increases. A randomly repeated local search algorithm is proposed which outperforms the other algorithms. The optimization scheme was applied to find the parameters based on tests with varying stress paths on the same soil by a combined parallelized optimization scheme. The combined optimization allows to find parameters that fit all stress paths as good as possible, serving as a basis for comparison of models.

The evaluation of the seismic response of underground structures is an important step in its design. Currently, displacement-based pseudo static methods are one of the most popular class of methods recommended by several design guidelines and codes. However, in this study, through comparisons with dynamic finite element method analysis using advanced soil constitutive models, we show that the application of the simplified displacement-based method in liquefiable soils is inappropriate. The seismic response of a typical rectangular section underground structure in non-liquefiable and liquefiable ground using linear elastic model and the unified constitutive model for large post-liquefaction deformation of sand, respectively. A displacement-based pseudo static method is then used in these two scenarios, using input from free field analysis. Comparisons between the results from the two types of analysis methods show relatively good agreement for the non-liquefiable ground scenario. Whereas for the liquefiable ground case, the displacement-based method significantly underestimates the deformation and internal forces of the structure. This is shown to be caused by the ineptness of the displacement-based pseudo static method in taking into consideration the spatial non-uniformity of structure and near field soil seismic response in liquefiable ground. This study highlights the importance of advanced constitutive modeling of soils in practical geotechnical earthquake engineering applications.

This paper presents preliminary results from the numerical simulation of cone penetration in Tresca material. The simulations were performed using two variations of the generalized interpolation material point method (GIMP), uniform GIMP and contiguous particle GIMP. A moving-compressible structured irregular background grid with adaptive material point splitting was used for computational efficiency. The non-linear B-bar method was used for mitigating volumetric locking. The obtained results were compared with results published in the literature to illustrate the efficacy and accuracy of using the proposed numerical scheme.

The material point method is ideally suited to modelling large deformation problems in three dimensions, especially in cases where the finite element method struggles due to mesh distortion. However, when the method is used to analyse problems with near-incompressible material behaviour, such as in geotechnical engineering using models with isochoric plastic flow, it suffers from severe volumetric locking. This causes the method to over predict the forces for a given displacement and induces spurious stress oscillations through the problem domain. Several methods have been proposed in the finite element literature but few of these have been applied to the material point method. In this paper we present a way to avoid volumetric locking for three-dimensional material point analyses with simplex elements (linear tetrahedra) using an $$ F $$ F bar patch approach. Not only does the technique avoid the over-stiff behaviour associated with volumetric locking but it also reduces the stress oscillations in the method, which are often attributed to cell-crossing instabilities. The formulation is validated against two three-dimensional benchmark problems.

A dynamic hydro-mechanical coupling analysis method using Material Point Method (MPM) and its applicability to liquefaction-induced large deformation problem are presented. Based on the u-p formulation, the governing equations comprise the equations of motion for the whole mixture and the continuity equation for the water phase, disretized by MPM and finite difference method, respectively. A cyclic elasto-plastic constitutive model is employed for simulating liquefaction behavior. As a numerical example, a river embankment mounted on liquefiable foundation ground is simulated using the MPM. Finite element analysis using the same governing equations and the constitutive mode is also presented for comparison. Evolution of pore water pressure and deformation mode are discussed. MPM more reasonably simulated large settlement and lateral flow of the river embankment and the flood channel induced by liquefaction than finite element method. In particular, MPM properly evaluated the large deformation-induced total stress changes which had a significant influence on the liquefaction behavior and subsequent deformation behaviors.

Anchors are popular and economic solutions for onshore and offshore engineering to provide adequate resistance for structures. In anchor design, it is critical to evaluate the bearing capacity of soil in supporting an anchor and understand the underlying load-carrying and failure mechanisms. Herein we employ a newly developed multiscale modeling tool based on coupled MPM/DEM for large deformation problems to investigate the cross-scale behavior during the pullout of plate anchors. This multiscale approach invokes Material Point Method (MPM) to tackle the large deformation problems in macro-scale and employs Discrete Element Method (DEM) to reproduce the soil response in meso-scale. It helps bypass conventional phenomenological constitutive models while providing cross-scale information of any interesting macroscopic observations. In this study, the influence of relative density of sand and the embedment ratio on the behavior of anchors (e.g., bearing capacity of anchors, deformation patterns) is investigated by the coupled method. Mesoscopic data and analyses originated from particle scale interactions are provided in lieu of the variable failure patterns observed in soil around an anchor during the pull-out process.

The material point method (MPM) has recently been used to perform soil–water interaction analysis for fully saturated soil in geotechnical engineering. In MPM, to treat fully saturated soil, two types of discretization methods have been mainly used: the single point formulation and the double point formulation. The former discretizes the soil–water mixture with a single set of particles, whereas the latter discretizes the soil phase and the water phase separately by using different sets of particles. In both the discretization methods, the flux boundary condition is imposed on the background mesh nodes, and the prescribed pore water pressure boundary condition is directly applied to the particles. This is the reason behind the complexity involved in the setting of boundary conditions for the coupled MPM. To reduce this complexity, this paper presents a discretization technique that combines MPM with smoothed particle hydrodynamics (SPH). SPH is used for the discretization of the generalized Darcy’s law to impose the flux boundary condition directly on the particles. The proposed formulation is validated through the soil–water interaction benchmark problems: the consolidation problems of Terzaghi and Mandel–Cryer. The numerical results are in good agreement with the analytical solutions, which validates the proposed method.

There is an increasing interest to utilise suction caissons as foundations for offshore wind turbines. Significant research has been devoted to developing penetration prediction methods and to understand the in-service response under cyclic loading. However, the effect of the installation process on the state of the surrounding soil is less well understood, although it may affect the in-service performance, in particular under relatively low magnitude cyclic loading, which represents the majority of loading conditions experienced by an offshore wind turbine in the field. This is due to the complexity in modelling the problem, which includes very large deformations, seepage flow and soil-structure interaction. Novel approaches featuring the material point method and centrifuge test results evaluated with the particle image velocimetry post analysis are capable of visualising the mechanisms underlying suction caisson installation. The results aim to reduce existing uncertainties and provide confidence in suction caissons as a reliable foundation system for offshore wind applications.

Soil-wheel interaction has been one of the fundamental research subjects in the terramechanics field. In this study the material point method (MPM) formulation is extended for the performance and mobility of unmanned wheels rotating on the soil. A new contact approach is developed for fast contact detection between rigid wheel and deformable soil. In each time step, contact background nodes are detected to define the contact elements and subsequently to determine the contact soil material points. An equivalent contact length is associated to each contact soil point and soil-wheel reaction forces are calculated. For a given angular velocity, the dynamic momentum balances (linear and angular) are solved and the motion of the wheel points is updated accordingly. Numerical simulations of a typical wheel going over the cohesive soil are presented to demonstrate the capabilities of the new approach. The soil is simulated using Mohr-Coulomb constitutive model, and the wheel is assumed as solid rigid. The wheel can automatically rotate and move forward vertically according to the material properties and surface geometry. FEM simulations is also conducted to compare and validate the MPM simulation, showing that the proposed MPM wheel-soil model is effective in predicting the dynamic rotation problems.

The material point method (MPM) shows promise for the simulation of large deformations in history-dependent materials such as soils. However, in general, it suffers from oscillations and inaccuracies due to its use of numerical integration and stress recovery at non-ideal locations. The development of a hydro-mechanical model, which does not suffer from oscillations is presented, including a number of benchmarks which prove its accuracy, robustness and numerical convergence. In this study, particular attention has been paid to the formulation of two-phase coupled material point method and the mitigation of volumetric locking caused numerical instability when using low-order finite elements for (nearly) incompressible problems. The numerical results show that the generalized interpolation material point (GIMP) method with selective reduced integration (SRI), patch recovery and composite material point method (CMPM) (named as GC-SRI-patch) is able to capture key processes such as pore pressure build-up and consolidation.

We focus on modeling geogenic arsenic mobilization induced by Fe-oxides reductive dissolution and aim at a preliminary assessment of the potential of such a mechanism to release arsenic under conditions that are typically found in alluvial aquifers. Under reducing conditions, Fe-oxides might be dissolved by iron-reducing bacteria, thus leading to simultaneous release of arsenic which is then typically absorbed and co-precipitated on Fe-oxides. We leverage on a kinetic model to describe this mechanism and resort to a global sensitivity analysis framework to explore the way the concentration of arsenic resulting from sediment-water interaction in a batch system is influenced by (i) available initial mass (per mass of sediments) of Fe-oxides, (ii) amount of arsenic initially adsorbed on Fe-oxides, and (iii) concentration of dissolved organic carbon for bacterial metabolism. Our analyses are conducive to (i) a preliminary assessment of the relative importance of multiple factors on the release of arsenic to groundwater by iron-oxides reductive dissolution and (ii) useful indications to improve the proposed geochemical model and its future transferability to flowing systems.

Uncertainty exists in geomaterials at contact, microstructural, and continuum scales. To develop predictive, robust multi-scale models for geotechnical problems, the new challenge is to allow for the propagation of model/parameter uncertainty (conditioned on laboratory/field measurements) between micro and macro scales. We aim to first quantify these uncertainties using an iterative Bayesian filtering framework. The framework utilizes the recursive Bayes’ rule to quantify the evolution of parameter uncertainties over time, and the nonparametric Gaussian mixture model to iteratively resample parameter space. Using the iteratively trained mixture to guide resampling, model evaluations are allocated asymptotically close to posterior modes, thus greatly reducing the computation cost. In this paper, we first respectively quantify the parameter uncertainty of models that are discrete and continuum in nature, namely a discrete particle and an elasto-plastic model. We then link the two models by conditioning their uncertainties on the same stress-strain response, thereby revealing micro-macro parameter correlations and their uncertainties. The micro-macro correlations obtained can be either general for any granular materials that share similar polydispersity or conditioned on the laboratory data of specific ones.

The contribution is motivated by the Bayesian approach to the solution of material identification problems which frequently appear in geo-engineering. We shall consider the cases with associated forward model describing flow in porous media with or without fractures as well as coupled hydro-mechanical processes. When assuming uncertainties in observed data, the use of the Bayesian inversion is natural. In comparison to deterministic methods, which lead only to a point estimate of the identified parameters, the Bayesian approach provides their probability distribution. The implementation of the Bayesian inversion is realized via Markov Chain Monte Carlo methods. The paper aims at the acceleration of the posterior sampling using a surrogate model that provides a polynomial approximation of the full forward model. The sampling procedure is based on the delayed acceptance Metropolis-Hastings (DAMH) algorithm. Therefore, for each proposed sample, the acceptance decision contains a preliminary step, which works only with an approximated posterior distribution constructed using the surrogate model. Furthermore, the approximated posterior distribution is being updated using new snapshots obtained during the sampling process. The posterior distribution updates are realized via updates of the surrogate model. The application of the described approach is shown through several model examples including flow in porous media with fractures and hydro-mechanical coupling.

In contrast to many disciplines, the approach to design in rock engineering remains largely inductive: observations, experience and engineering judgment are used to infer the behavior of a problem that cannot be constrained due to the nature of geological/geotechnical materials. Most of the rock mass classification systems used for rock engineering design purposes were developed in the 1960s and 1970s; since then no major updates have been proposed to reflect modern data collection tools and modelling procedures. Furthermore, engineers have attempted to apply existing classification systems in the context of a probabilistic design approach despite most of those systems being based on qualitative and semi-quantitative measurements. In this paper, we use a discrete fracture network (DFN) approach to introduce the first component of a new quantitative classification system that can capture rock mass scale, anisotropic effects; and better reflects the degree of connectivity of the natural fracture network. A new network connectivity index (NCI) is introduced that uses areal fracture intensity and density, and intersection density to provide a quantitative description of rock mass blockiness.

The use of Data Assimilation (DA) techniques is receiving an increasing interest in geomechanical applications, with the aim to assess and reduce uncertainties associated to numerical outcomes by model constrain with available measurements. In geomechanical simulations, ensemble-based DA approaches are usually preferred. Among such techniques, Ensemble Smoother with Multiple Data Assimilation (MDA-ES) is usually recognized to improve the outcomes in nonlinear problems, but its use and parameter definition is still object of research. In this paper, MDA-ES has been tested in a synthetic case study dealing with the prediction of land subsidence above a producing hydrocarbon reservoir. Its effectiveness has been investigated varying both the DA parametrization and the geomechanical properties.

A numerical analysis was validated for shear and consolidation deformation of clay ground base on uncertainty quantification referring ASME V&V10.1 (2012). The numerical analysis is a soil-water coupled analysis on finite deformation porous media theory. First, centrifugal experiments with the established conditions were performed, and probability distributions of deformation of clay ground were obtained. Second, some triaxial tests with the established conditions were performed, and some elasto-plastic material parameters were assumed as normal distributions from the test results. Third, numerical analyses simulated the centrifugal experiments of clay ground deformation using the uncertainty of elasto-plastic parameters and the embankment load. By comparing the experimental and numerical results, the uncertainty quantification of clay ground deformation and validity of numerical method were discussed. The area matric MSRQ was used to assess the validity. In this example, MSRQ of vertical displacement at the embankment toe was quite large due to the biased distribution of numerical solutions.

Application of pipe-jacking method in the form of microtunneling has become more popular over the conventional open cut method for the installation of underground infrastructure such as buried sewer pipelines in urban setting in recent years. This is due to the advantages offered by trenchless technology such as reduced disruptions to traffic and the surrounding environment as well as minimized ground settlements. Prediction of frictional jacking forces is a crucial component of the design of pipe-jacking works. In view of the challenges faced in calculating pipe-jacking forces in highly weathered and highly fractured geological formations, this paper proposes the use of Bayesian inference method to predict the frictional jacking forces developed from traversing the weathered rock formations. A probabilistic framework based on Bayesian approach is proposed using a well-established pipe-jacking force model, which considers arching effect from the surrounding ground. The main advantages of Bayesian inference include (i) consideration of uncertainty in deriving the soil parameters and (ii) ability to incorporate prior information and expert judgement from previous research studies into the model in the form of prior distribution. The model uncertainty is expected to be significantly reduced through the sequential updating process when more data become available.

The detection of piping and the extent to which it progresses is important in the serviceability of natural dams and levees, where the soil is usually poorly consolidated and filters or drainage zones are absent. Practically, information of the spatial variation of hydraulic conductivity is also unavailable in the domain of interest. An approximate version of a Hamiltonian Monte Carlo (HMC) based method for combined probabilistic inversion is detailed. The inverse problem consists of Karhunen-Loève (KL) expansion parameters and solid-void interface parameters which are determined simultaneously. The interface parameter updates are carried out using a reversible proposal in a mesh moving framework. To maintain computational efficiency, the parameter update is enforced in an approximate sense, where the covariance matrix for the KL expansion is kept constant throughout the analysis. Synthetic data from a numerical experiment on a domain containing a predefined piping region, is used to validate the approximate method. A total of 30000 samples are generated using the HMC sampler. Results show that the Markov chains converge to the stationary distribution. A good match is also observed between the inferred mean interface, and the true interface and the true spatial distribution of hydraulic conductivity is obtained.

Rock failure depends on a high number of variables, one of which is the distribution of defects. To estimate the effect of internal defects on strength variability, numerical simulations of mechanical tests under different loading configurations were run on samples where specific defect distributions were implemented. The Discrete Element Method (DEM) was used to generate three-dimensional rock samples using a bonded particle model. The spheres were joined using cohesive bonds with an interaction radius that allows non contacting spheres within this interaction radius to interact together. To simulate the defect distribution inside rock samples, an internal Weibull distribution of the bond strength was applied. Multiple standard laboratory tests generally used for material characterization were simulated. The Weibull distribution shape parameter of the strength distribution of the samples was then computed. The resulting shape parameters were used to establish a correlation between the shape parameter of the internal strength distribution and the macroscopic shape parameter of the strength distribution.

The excavation and construction of metro structures in a high-density urban area require cautions in the design process. The knowledge of ground should warn the engineers and enable them to take measures to avoid or at least to minimize the potential risks. To achieve that task, numerical simulations are used to determine the deformations due to the construction process. The role played by soil behavior is as fundamental as the values of its parameters. Due to the heterogeneity character of the soil, its random parameters influence on the response of the numerical simulation. A good understanding of that response could be addressed through sensitivity analysis, that is, the assessment of the impact of individual input parameters or sets of input parameters on the response of the model. In this paper, a sensitivity analysis is performed to illustrate the effect of the uncertainty of the input parameters on the response of a numerical simulation of tunnel construction through the finite element code Zsoil. For this purpose, the first-order sensitivity and the total order sensitivity based on Fourier Amplitude Sensitivity (FAST) technics will be performed.

The study of micromechanical behaviour of granular materials has a continuous interest in several engineering fields. Better understanding of the relationship between particle morphology and mechanical performance of such materials is essential in geotechnical applications. Past studies have used discrete element modelling (DEM) to demonstrate the particulate behaviour of granular soils. However, the material properties used as input parameters for DEM simulations are often generalised due to limited data used during calibration and validation exercises. With the advancement of micro-CT scan and 3D printing, improvement can be made to DEM simulations through independent study of particle response for various materials. This paper describes the methodology used to improve DEM simulations and to replicate granular particles by means of 3D printing. Realistic geometric extraction of sand particles was achieved through micro-CT which can then be imported to DEM simulations. To study the influence of particle mineralogy, synthetic particles were produced by means of 3D printing. A range of sintering powders can be used to print particles with various stiffness. The outcomes from the improved DEM models and testing of synthetic printed particles can be used to validate the mechanical behaviour and particle interaction through the independent study of particle morphology, size, angularity, gradation and mineralogy.

Field measurements can be used to improve the estimation of the performance of geotechnical projects (e.g. embankment slopes, soil excavation pits). Previous research has utilised inverse analysis (e.g. the ensemble Kalman filter (EnKF)) to reduce the uncertainty of soil parameters, when measurements are related to the performance, such as inflow, hydraulic head, deformation, etc. In addition, there are also direct measurements, such as CPT measurements, where parameters (i.e. tip resistance and sleeve friction) can be directly correlated with, e.g. soil deformation and/or strength parameters, where conditional simulation via constrained random fields can be used to improve the estimation of the spatial distribution of parameters. This paper combines these two (i.e. direct and indirect) methods together in a soil excavation analysis. The results demonstrate that the parameter uncertainty (and thereby the uncertainty in the response) can be significantly reduced when the two methods are combined.

It is often of great interest in geomechanical analysis to model localized phenomena, such as strain concentrations along failure planes or hydraulic gradient concentrations near discontinuities and exit points. Focus in recent years on massively parallel finite element codes with adaptive meshing algorithms, have left many previously developed libraries for adapting quadrilateral meshes difficult to find, and often unsupported. While large, parallel simulations are of great interest to researchers, an acceptable modeling outcome for many practical engineering problems can often be obtained using meshes with more modest numbers of elements (e.g. < 105), removing the need for use of large parallel libraries. This paper describes a software library written in modern Fortran for adaptive meshing of structured meshes. The library consists of less than 1,000 lines of code and is based on 4 Fortran functions making it quick to learn as an educational and research tool, and easy to implement in existing codes. The library is demonstrated through an example seepage problem of a point source (Poisson problem).

Dynamic compaction is a process for densifying soils by repeatedly dropping a large tamper from a crane, and this process can be applied to densify a wide range of soils in place to depths greater than ten metres. This procedure has proved to be useful in reducing the potential for settlements associated with wetting of unsaturated soils. Although field experience and experimental tests indicate that compaction efficiency is related to the soil moisture content, the compaction process makes it difficult to install some measurement instruments below the impact areas to study different aspects of the response of unsaturated soils, such as suction and degree of saturation changes. Therefore, numerical modelling could be a more cost-effective way to gain a better understanding of the dynamic response of unsaturated soils and ultimately to improve the efficiency of dynamic compaction on unsaturated soils with different water contents. Due to the numerical difficulties caused by the complexity of constitutive models of unsaturated soils and the dynamic response of a three-phase system, the numerical simulation of dynamic compaction of soils is necessarily limited. In this paper, a finite element model has been developed with an advanced constitutive model implemented, and the generalised-α method has been applied to solve the global equations of motion.

Numerical modeling of anthropogenic land subsidence due to the exploitation of subsurface resources is of major interest to anticipate possible environmental impacts on the ground surface. The reliability of predictions depends on different sources of uncertainty introduced into the modeling procedure. In this study, we focus on reduction of model parameter uncertainty via assimilation of land surface displacements. A test case application on a deep hydrocarbon reservoir is considered where land settlements are predicted with the aid of a 3D Finite Element (FE) model. The calibration of the parameters defining the rock constitutive law is obtained by the Ensemble Smoother (ES) technique. The ES convergence is guaranteed with a large number of Monte Carlo simulations that may be computationally infeasible in large scale and complex systems. A surrogate model based on the generalized Polynomial Chaos Expansion (gPCE) is proposed as an approximation of the forward problem. This approach is expected to reduce the overall computational cost of the original ES formulation and enhance the accuracy of the parameter estimation problem. The result is compared with a posterior sampling by Markov Chain Monte Carlo (MCMC) to assess the quality of the assimilation.

Contact problems are widely encountered in geotechnical engineering, such as the contact between soils and the concrete slab in earth and rockfill dams, concrete lining in a tunnel and coastal levees. Due to the unknown contact region and contact forces, the contact problems have strong boundary nonlinearity. In addition, soils have been recognized as heterogeneous materials in geotechnical engineering, and the existence of the spatial variability of soils increases the nonlinearity of the problems. In order to investigate the influence of heterogeneity on the contact problems, the penalty method is used to analyse the contact problems. In this paper, Young’s modulus, is taken to be a spatially variable. Random field theory is used to model the heterogeneity of Young’s Modulus. The results showed that the influence of heterogeneity on the elastic contact problems is significant. In order to better predict the deformation/stress in the contact bodies, the spatial variability needs to be considered.

Numbers of uncertainties are involved in slope stability analysis. In particular, when a slope fails, uncertainty of parameters of the slope will cause the variability of runout distance of sliding mass which decides the consequence triggered by landslides. In general, slope geometry parameters and residual strengths parameters are highly close to the runout distance of landslides. This paper aims to investigate the effect of residual strength parameters and initial geometry parameters on runout distance of landslides via material point method. To clearly indicate relation of runout distance and these parameters for different soil slopes, three kinds of slopes, i.e., slopes in single soil layer, slopes in weaker upper soil layer and stronger lower soil layer and slopes in stronger upper soil layer and weaker lower soil layer are studied. For the first and third kinds of slopes, it is found that runout distance has a good negative exponential law with residual strength parameters and has a significant linear increase trend with geometry parameters. For the second kind of slope, the decrease trend of runout distance with residual strength parameters is faster; The increase trend of the runout distance with slope gradient can be indicated well via sigmoid curve; There is also a good linear increase relation of the runout distance and slope elevation. Additionally, the relation between runout distance and slope elevation obtained via MPM in this paper is consistent with that based on the empirical method. The conclusions can provide reference for assessment and management of landslides risk.

A probabilistic analysis framework in which the spatial variability and the statistical uncertainty are considered simultaneously is presented to evaluate an overall strength of a cement-treated soil column. In this framework, the statistical uncertainty is evaluated from sample data using a Bayesian inference. The inference about the statistical parameters is performed by a Markov chain Monte Carlo method. The drawn values of the parameters are adopted when making random fields of strength as realizations. Then, the finite element analysis is conducted for the generated realizations. The framework components are briefly described, and an example analysis is performed to illustrate the influence of the statistical uncertainty and the spatial variability on the evaluation of the overall strength.

The construction of a geotechnical model for a tunnel is a challenging process, often complicated by the limited quantity of available data compared to the extension of the project. In this process two main activities concur to determine the result: i) the geotechnical characterization and ii) the geological characterization. The first one derives the materials properties, the second allows the extrapolation of these properties at the large scale. In this paper, after some preliminary considerations on the interplays existing between these two activities, the attention is principally focused on the second one, that is, on the construction of a geological model and on the possibility of obtaining a quantitative estimate of its reliability. Elements contributing to the capacity of the model of providing appropriate forecasts are considered and their mutual interactions are explored. A multistep computation process is illustrated leading to the derivation of an index called GMR, whose dimension is related to the reliability of the geological model.

Modern engineering problems are facing the growing demand to deal with huge amount of data and their intrinsic uncertainties. This exigence has led us to unprecedented insights and developments in the machine learning field. To date, the healthcare and financial sectors has been the precursor of practical application of machine learning approaches. In geotechnics and rock mechanics, the materials we deal with are characterized by a large amount of data, various levels of uncertainty and often a prior knowledge, therefore they lend themselves well to this type of analysis. This article aims to present Bayesian methods and machine learning algorithms applied for geotechnical characterization of soil and rocks. Once the test sample has been properly filtered and classified, we will demonstrate the potentiality of multivariate Bayesian linear regression as a main tool for dealing with multivariate data and uncertainty. In addition to frequentist approaches, we will make use of Bayesian models where the regression parameters, based on a prior distribution, will be calculated in terms of mean and variance.

In densely populated areas, it is common the need for using deep excavations for different infrastructure works, but they may cause excessive ground movements and generate damages to neighboring buildings. The analysis of deep excavations in soft soils is a complex problem in which the use of a combination of numerical and probabilistic techniques is useful to represent the behavior of such inherently variable deposits. In the numerical and probabilistic analysis, constitutive soil parameters can be treated as random fields to deal with uncertainty due to spatial variability. In this research, parameters E50ref and E0ref from the constitutive model Hardening Soil Small Strain are simulated as random fields. Numerical analyses are performed on tridimensional models of deep excavations in Bogotá soft soils, obtaining system response in terms of damage potential indexes in adjacent buildings. The damage probabilities are assessed for each building and simulated construction stage. Obtained damage probabilities from random field-based modeling are compared with the ones obtained from random-variable based modeling.

This paper shows how accurately and efficiently reliability analyses of geotechnical installations can be performed by directly coupling geotechnical software with a reliability solver. An earth slope is used as the study object. The limit equilibrium method of Morgenstern-Price is used to calculate factors of safety and find the critical slip surface. The deterministic software package Slope/w is coupled with the StRAnD reliability software. Reliability indexes of critical probabilistic surfaces are evaluated by the first-order reliability methods (FORM). By means of sensitivity analysis, the effective cohesion ( $$ c^{\prime} $$ c ′ ) is found to be the most relevant uncertain geotechnical parameter for slope equilibrium. The slope was tested using different geometries. Finally, a critical slip surface, identified in terms of minimum factor of safety, is shown here not to be the critical surface in terms of reliability index.

Groundwater abstractions may lead to structural failure of buildings due to the settlement of the subsurface. In order to set the probability of failure to a predefined acceptable level, a probabilistic procedure is proposed in this article. The procedure provides intervention levels for groundwater heads at which the extraction of groundwater should be reduced. As part of the procedure a Levenberg-Marquardt method, extended by a subspace regularization technique, calibrates a finite element method based groundwater model that simulates flow under natural conditions or at moderate extraction rates. The method provides expected values for geo-hydrologic parameters like transmissivity and flow resistance and expresses the parameter uncertainty in a covariance matrix. A first order reliability method uses these values and calculates the failure probability conditioned for a series of intervention levels. This paper outlines the procedure and presents the results of a verification test. A pumping test near Utrecht in the Netherlands is used to simulate a field scale application. The application shows that the procedure reduces the risk of failure to an acceptable level.