Dilatancy-controlled gas flow in preferential pathways plays a key role in the safety analysis of radioactive waste repositories. This is particularly the case for bentonite, an often-preferred barrier material. Gas flow in preferential pathways is characterized by localization and spontaneous behavior, which is challenging to simulate in numerical models due to strong hydro-mechanical coupling. To analyze a laboratory experiment in the framework of the DECOVALEX-2023 project, this study introduced a new approach of combining continuous modelling methods with spatial material properties derived from material heterogeneities and experimental observations. The proposed model utilized hydro-mechanical spatial distributions, namely Young's modulus and gas entry pressure, and elastoplasticity combined with a linear swelling model. A conceptual strain-dependent permeability approach simulated dilatancy-controlled gas flow based on hydro-mechanical coupling. To test the effectiveness of the presented approach, a gas injection test in a compacted, saturated bentonite sample was simulated using the open-source code OpenGeoSys 5.8 and compared with experimental observations. The presented methodology is capable of simulating localized gas flow in preferential pathways. The spatial distributions of Young's modulus and gas entry pressure affect the swelling pressure, relative permeability and, in combination with the strain-dependent permeability model, also the intrinsic permeability.

Carbonate rocks are extensively used in civil infrastructure and play a critical role in geoenergy geoengineering, either as hydrocarbon reservoirs or potential repositories for CO2 geological storage. Carbonate genesis and diagenetic overprint determine the properties of carbonate rocks. This study combines recent data gathered from Madison Limestone and an extensive dataset compiled from published sources to analyze the hydraulic and mechanical properties of limestone carbonate rocks. Physical models and data analyses recognize the inherently granular genesis of carbonate rocks and explain the strong dependency of physical properties on porosity. The asymptotically-correct power model in terms of (1-ϕ/ϕ∗)α is a good approximation to global trends of unconfined stiffness E and unconfined compressive strength UCS, cohesive intercept in Mohr-Coulomb failure envelopes, and the brittle-to-ductile transition stress. This power model is the analytical solution for the mechanical properties of percolating granular structures. We adopted a limiting granular porosity ϕ∗ = 0.5 for all models, which was consistent with the loosest packing of monosize spheres. The fitted power model has exponent (α = 2) in agreement with percolation theory and highlights the sensitivity of mechanical properties to porosity. Data and models confirm a porosity-independent ratio between unconfined stiffness and strength, and the ratio follows a log-normal distribution with mean (E/UCS) ≈ 300. The high angle of internal shear strength measured for carbonate rocks reflects delayed contact failure with increased confinement, and it is not sensitive to porosity. Permeability spans more than six orders of magnitude. Grain size controls pore size and determines the reference permeability k∗ at the limiting porosity ϕ∗ = 0.5. For a given grain size from fine to coarse-grained dominant carbonates, permeability is very sensitive to changes in porosity, suggesting preferential changes in the internal pore network during compaction.

This study proposed an analytical model for the tunnel supported with a tangentially yielding liner in viscoelastic ground. The efficiency of the developed analytical model was verified by comparing the calculated results with associated numerical simulation results. Using the analytical model, a comprehensive parameter sensitivity analysis was performed to examine the effects of the rate of tunnel face advancement, concrete liner thickness, installation time of liner, and strength and thickness of yielding elements on the tunnel responses. The results highlight the significant benefit of the tangentially yielding liner to relieve overstress in the tunnel liner and improve the stability of the tunnel. The yield efficiency of the tangentially yielding liner depends highly on the yielding strength and deformable capacity of the yielding elements and less on the installation time.

Quantitative damage identification of surrounding rock is important to assess the current condition and residual strength of underground tunnels. In this work, an underground tunnel model with marble-like cementitious materials was first fabricated using the three-dimensional (3D) printing technique and then loaded to simulate its failure mode in the laboratory. Lead zirconate titanate piezoelectric (PZT) transducers were embedded in the surrounding rock around the tunnel in the process of 3D printing. A 3D monitoring network was formed to locate damage areas and evaluate damage extent during loading. Results show that as the load increased, main cracks firstly appeared above the tunnel roof and below the floor, and then they coalesced into the tunnel boundary. Finally, the tunnel model was broken into several parts. The resonant frequency and the peak of the conductance signature firstly shifted rightwards with loading due to the sealing of microcracks, and then shifted backwards after new cracks appeared. An overall increase in the root-mean-square deviation (RMSD) calculated from conductance signatures of all the PZT transducers was observed as the load (damage) increased. Damage-dependent equivalent stiffness parameters (ESPs) were calculated from the real and imaginary signatures of each PZT at different damage states. Satisfactory agreement between equivalent and experimental ESP values was achieved. Also, the relationship between the change of the ESP and the residual strength was obtained. The method paves the way for damage identification and residual strength estimation of other 3D printed structures in civil engineering.

The rock fracture characteristics and principal stress directions are crucial for prevention of geological disasters. In this study, we carried out biaxial compression tests on cubic granite samples of 100 mm in side length with different intermediate principal stress gradients in combination with acoustic emission (AE) technique. Results show that the fracture characteristics of granite samples change from ‘sudden and aggregated’ to ‘continuous and dispersed’ with the increase of the intermediate principal stress. The effect of increasing intermediate principal stress on AE amplitude is not significant, but it increases the proportions of high-frequency AE signals and shear cracks, which in turn increases the possibility of unstable rock failure. The difference of stress in different directions causes the anisotropy of rock fracture and thus leads to the obvious anisotropic characteristics of wave velocity variations. The anisotropy of wave velocity variations with stress difference is probable to identify the principal stress directions. The AE characteristics and the anisotropy of wave velocity variations of granite under two-dimensional stress are not only beneficial complements for rock fracture characteristic and principal stress direction identification, but also can provide a new analysis method for stability monitoring in practical rock engineering.

To investigate the influence of unloading effect of a circular tunnel face on rockburst process, by innovatively combining rock drilling unloading devices and triaxial systems, the strain rockburst simulation under the entire stress path of “high initial stress + internal unloading + stress adjustment” (HUS test) was realized for the intact cubic red sandstone samples (100 mm × 100 mm × 100 mm). Comparative tests were conducted on cubic red sandstone samples with prefabricated circular holes (φ 25 mm) under the stress path of “prefabricated circular hole + high initial stress + stress adjustment” (PHS test), thereby highlighting the influence of internal unloading on rockburst failure. The test results revealed that with an increase in vertical stress, the sidewalls in both the HUS and PHS tests suffered strain rockburst failure. Compared with the PHS test, the initial failure stress in the HUS test is lower, and it is easier to induce sidewall rockbursts. This indicates that the internal unloading influences the sidewall failure, causing an obvious strength-weakening effect, which becomes more significant with an increase in buried depth. The strain rockburst failure was more severe in the HUS test owing to the influence of internal unloading. V-shaped rockburst pits were formed in the HUS tests, whereas in the PHS test, arc-shaped rockburst pits were produced. It was also found that strain rockburst failure may occur only when the rock has a certain degree of rockburst proneness.

Predicting the mechanical behaviors of structure and perceiving the anomalies in advance are essential to ensuring the safe operation of infrastructures in the long run. In addition to the incomplete consideration of influencing factors, the prediction time scale of existing studies is rough. Therefore, this study focuses on the development of a real-time prediction model by coupling the spatio-temporal correlation with external load through autoencoder network (ATENet) based on structural health monitoring (SHM) data. An autoencoder mechanism is performed to acquire the high-level representation of raw monitoring data at different spatial positions, and the recurrent neural network is applied to understanding the temporal correlation from the time series. Then, the obtained temporal-spatial information is coupled with dynamic loads through a fully connected layer to predict structural performance in next 12 h. As a case study, the proposed model is formulated on the SHM data collected from a representative underwater shield tunnel. The robustness study is carried out to verify the reliability and the prediction capability of the proposed model. Finally, the ATENet model is compared with some typical models, and the results indicate that it has the best performance. ATENet model is of great value to predict the real-time evolution trend of tunnel structure.

In karst areas, the drainage pipes of aging tunnels are prone to be clogged by precipitated carbonates, resulting in lining cracking and tunnel leaking. As a result, not only the driving safety will be deteriorated, but also the water pressure on the lining might also be elevated significantly. For the structural stability and service lifespan of old tunnels, it is of great importance to remove these precipitated carbonates in time. Traditional treatment methods are often destructive to some extent or not efficient enough. This study aims to experimentally develop an eco-friendly acid-based chemical cleaning method to remove carbonate precipitations efficiently. The proposed chemical cleaning agent is an aqueous solution with strong acidity, consisting of sulfamic acid, water, and additives. The factors affecting the cleaning efficiency include the acid solubility, temperature and flow rate of the cleaning agent, as well as additives. Elevating the solution temperature to 50 °C or a flow rate of no less than 0.2 m/s can improve cleaning efficiency. Although the salt effect cannot work, 1 wt% of polymaleic acid as a surfactant could further promote the cleaning rate. The cleaning efficiency will increase with the flow rate in a power function. The relatively low flow rate that improves the cleaning rate considerably can avoid high-pressure-induced mechanical damage to tunnel drainpipes. The waste could be easily treated to acceptable levels using commercial sewage treatment products and can also be recycled in agriculture. With the chemical cleaning, the water pressure at the arch springing of the lining will reduce with the increased radius of transverse drainpipes in a power function. The proposed acid-based cleaning method, which is highly efficient, non- or low-destructive to aging tunnels, sufficiently safe for humans, and friendly enough to the environment, will offer a promising alternative to remove the precipitated carbonates in tunnel drainpipes efficiently.

Rock mass is a fractured porous medium usually subjected to complex geostress and fluid pressure simultaneously. Moreover, the properties of rock mass change in time and space due to mining-induced fractures. Therefore, it is always challenging to accurately measure rock mass properties. In this study, a three-dimensional (3D) microseismic (MS) data-driven damage model for jointed rock mass under hydro-mechanical coupling conditions is proposed. It is a 3D finite element model that takes seepage, damage and stress field effects into account jointly. Multiple factors (i.e. joints, water and microseismicity) are used to optimize the rock mass mechanical parameters at different scales. The model is applied in Shirengou iron mine to study the damage evolution of rock mass and assess the crown pillar stability during the transition from open-pit to underground mining. It is found that the damage pattern is mostly controlled by the structure, water and rock mass parameters. The damage pattern is evidently different from the two-dimensional result and is more consistent with the field observations. This difference is caused by the MS-derived damage acting on the rock mass. MS data are responsible for gradually correcting the damage zone, changing the direction in which it expands, and promoting it to evolve close to reality. For the crown pillar, the proposed model yields a more trustworthy safety factor. In order to guarantee the stability of the pillar, it is suggested to take waterproof and reinforcement measures in areas with a high degree of damage.

Sand liquefaction under static and dynamic loading can cause failure of embankments, slopes, bridges and other important infrastructure. Sand liquefaction in the seabed can also cause submarine landslides and tsunamis. Fabric anisotropy related to the internal soil structure such as particle orientation, force network and void space is found to have profound influence on sand liquefaction. A constitutive model accounting for the effect of anisotropy on sand liquefaction is proposed. Evolution of fabric anisotropy during loading is considered according to the anisotropic critical state theory for sand. The model has been validated by extensive test results on Toyoura sand with different initial densities and stress states. The effect of sample preparation method on sand liquefaction is qualitatively analysed. The model has been used to investigate the response of a sand ground under earthquake loading. It is shown that sand with horizontal bedding plane has the highest resistance to liquefaction when the sand deposit is anisotropic, which is consistent with the centrifuge test results. The initial degree of fabric anisotropy has a more significant influence on the liquefaction resistance. Sand with more anisotropic fabric that can be caused by previous loading history or compaction methods has lower liquefaction resistance.

Local soil conditions can significantly modify the seismic motion expected on the soil surface. In most cases, the indications concerning the influence of the underlying soil provided by the in-force European and Italian Building Codes underestimate the real seismic amplification effects. For this reason, numerical analyses of the local seismic response (LSR) have been encouraged to estimate the soil filtering effects. These analyses are generally performed in free-field conditions, ignoring the presence of superstructures and, therefore, the effects of dynamic soil-structure interaction (DSSI). Moreover, many studies on DSSI are characterised by a sophisticated modelling of the structure and an approximate modelling of the soil (using springs and dashpots at the foundation level); while others are characterised by a sophisticated modelling of the soil and an approximate modelling of the structure (considered as a simple linear elastic structure or a single degree of freedom system). This paper presents a set of finite element method (FEM) analyses on a fully-coupled soil-structure system for a reinforced concrete building located in Fleri (Catania, Italy). The building, designed for gravity loads only, was severely damaged during the 26 December 2018 earthquake. The soil was modelled considering an equivalent visco-elastic behaviour, while the structure was modelled assuming both the visco-elastic and visco-inelastic behaviours. The comparison made between the results of the FEM analyses and the observed damage is valuable.

This study examines the effect of nanosilica (NS) additive to improve the mechanical properties of clay, clayey sand, and sand. The engineering properties of the soils were investigated through Atterberg limits, compaction, unconfined compression, ultrasonic pulse velocity (UPV), freeze-thaw, and direct shear tests. The NS content varied from 0% to 0.7% and cement content was 5% and 10% by the dry weight of the soil. The curing period varied from 7 d to 150 d. The consistency, compaction, and strength properties of the soils were affected by the presence of NS and cement. The optimum NS contents in clay specimens with 5% and 10% cement were 0.5% and 0.7%, respectively. It was 0.7% in sand specimens with both cement ratios, as well as 0.3% and 0.7% in clayey sand specimens with 5% and 10% cement, respectively. In terms of freeze-thaw resistance, clayey sand specimens containing 0.5% NS and 10% cement had the minimum strength loss. Exponential relationships existed between the ultrasonic pulse velocity (UPV) and the unconfined compressive strength (UCS) of soil specimens having the same curing period. The shear strength parameters of the soils also improved with the addition of NS. Scanning electron microscope (SEM) images demonstrated that cement and NS contributed to the improvement of the soils by producing a denser and more uniform structure. It was concluded that the minor addition of NS could potentially improve the geomechanical properties of the soils.

Levees are essential structures in flood defense systems, and their failures can lead to devastating consequences on the surrounding territories. One of the failure mechanisms mostly controlled by the foundation soil stratigraphy is the instability of the land side slope, triggered by the development of high uplift pressures in the foundation. This complex phenomenon has been investigated experimentally with centrifuge tests or large-scale tests and numerically with the limit equilibrium method (LEM) and the finite element method (FEM). In this work, we applied a multiphase formulation of the material point method (MPM) to analyze the development of toe uplift instability mechanism, from the onset of failure to large displacements. The numerical model is inspired by an experiment carried out in a geotechnical centrifuge test by Allersma and Rohe (2003). The comparison with the experiment allows for understanding critical pore pressure triggering large displacements in the foundation soils. Moreover, we numerically evaluated the impact of different values of foundation soils' hydraulic conductivity on the failure mechanism. The results show that hydraulic conductivity mainly influences the time of failure onset and the extension of shear localization at depth. Finally, the advantages of using large displacement approaches in the safety assessment of earth structures are discussed. Unlike FEM, there are no issues with element distortions generating difficulties with numerical convergence, allowing for full post-failure reproduction. This capability permits precise quantification of earth structure damages and post-failure displacements. The ensuing reinforcement systems' design is no longer over-conservative, with a significant reduction in associated costs.

Because of the cementation inherited from the parent rock, weathered granitic soil is usually susceptible to disturbance, which poses considerable challenges for laboratory characterization. The cone penetration test with pore pressure measurements has long been known for its reliability in site investigations and stratigraphic profiling. However, although extensive piezocone test results and experience are available for sedimentary soil, similar advances are yet to be made for weathered granitic soil. Moreover, the experience from sedimentary soil may not be directly applicable to weathered profiles because of the essentially different natures of the two types of geomaterials. This study performs seismic piezocone tests in a weathered granitic profile comprising residual granitic soil, completely weathered granite, and highly weathered granite. Pore pressure is measured at both the cone mid-face and the shoulder, and the effects of penetrometer size and penetration rate are considered. A series of updated soil behavior type charts is proposed to interpret the test results, thereby allowing the effect of weathering to be evaluated. This paper offers an important extension to the sparse data on the in situ responses of weathered materials.

The surge in demand for renewable energy to combat the ever-escalating climate crisis promotes development of the energy-saving, carbon saving and reduction technologies. Shallow ground-source heat pump (GSHP) system is a promising carbon reduction technology that can stably and effectively exploit subsurface geothermal energy by taking advantage of load-bearing structural elements as heat transfer medium. However, the transformation of conventional geo-structures (e.g. piles) into heat exchangers between the ground and superstructures can potentially induce variable thermal axial stresses and displacements in piles. Traditional energy pile analysis methods often rely on deterministic and homogeneous soil parameter profiles for investigating thermo-mechanical soil-structure interaction, without consideration of soil spatial variability, model uncertainty or statistical uncertainty associated with interpolation of soil parameter profiles from limited site-specific measurements. In this study, a random finite difference model (FDM) is proposed to investigate the thermo-mechanical load-transfer mechanism of energy piles in granular soils. Spatially varying soil parameter profile is interpreted from limited site-specific measurements using Bayesian compressive sensing (BCS) with proper considering of soil spatial variability and other uncertainties in the framework of Monte Carlo simulation (MCS). Performance of the proposed method is demonstrated using an illustrative example. Results indicate that the proposed method enables an accurate evaluation of thermally induced axial stress/displacement and variation in null point (NP) location with quantified uncertainty. A series of sensitivity analyses are also carried out to assess effects of the pile-superstructure stiffness and measurement data number on the performance of the proposed method, leading to useful insights.

A suitable bearing capacity of foundation is critical for the safety of civil structures. Sometimes foundation reinforcement is necessary and an effective and environmentally friendly method would be the preferred choice. In this study, the potential application of enzyme-induced carbonate precipitation (EICP) was investigated for reinforcing a 0.6 m bedding layer on top of clay to improve the bearing capacity of the foundation underneath an underground cable duct. Laboratory experiments were conducted to determine the optimal operational parameters for the extraction of crude urease liquid and optimal grain size range of sea sands to be used to construct the bedding layer. Field tests were planned based on orthogonal experimental design to study the factors that would significantly affect the bio-cementation effect on site. The dynamic deformation modulus, calcium carbonate content and long-term ground stress variations were used to evaluate the bio-cementation effect and the long-term performance of the EICP-treated bedding layer. The laboratory test results showed that the optimal duration for the extraction of crude urease liquid is 1 h and the optimal usage of soybean husk powder in urease extraction solution is 100 g/L. The calcium carbonate production rate decreases significantly when the concentration of cementation solution exceeds 0.5 mol/L. The results of site trial showed that the number of EICP treatments has the most significant impact on the effectiveness of EICP treatment and the highest dynamic deformation modulus (Evd) of EICP-treated bedding layer reached 50.55 MPa. The area with better bio-cementation effect was found to take higher ground stress which validates that the EICP treatment could improve the bearing capacity of foundation by reinforcing the bedding layer. The field trial described and the analysis introduced in this paper can provide a practical basis for applying EICP technology to the reinforcement of bedding layer in poor ground conditions.

Slope bearing capacity is one of the most important characteristics in slope engineering and is strongly influenced by weak planes, loading conditions, and slope geometry. By presenting the evolution of slip surfaces, this paper explored how the slope bearing capacity is affected by widely observed influencing factors. The initiation and propagation of slip surfaces are presented in laboratory model tests of slope using the transparent soil technique. Shear band evolution under various weak planes, loading conditions, and slope geometries were experimentally presented, and slope bearing capacities were analyzed with the process of shear band evolution. This paper verified that slip surface morphologies have a strong relation with the slope bearing capacity. The same slip surface morphology can have different evolutionary processes. In this case, it is the shear band evolution that determines the slope bearing capacity, not the morphology of the slip surface. The influencing factors such as pre-existing weak planes, loading conditions, and slope geometry strongly affect the slope bearing capacity as these factors govern the process of shear band evolution inside the slope.

Predicting the constitutive response of granular soils is a fundamental goal in geomechanics. This paper presents a machine learning (ML) framework for the prediction of the stress-strain behaviour and shear-induced contact fabric evolution of an idealised granular material subject to triaxial shearing. The ML-based framework is comprised of a set of mini-triaxial tests which provide a benchmark for the setup and validation of the discrete element method (DEM) model of the granular materials, a parametric DEM simulation programme of virtual triaxial tests which provides datasets of micro- and macro-mechanical information, as well as a multi-layer perceptron (MLP) neural network which is trained and tested using the DEM-based datasets. The ML model only requires the initial void ratio of the granular sample as the input for predicting its constitutive response. The excellent agreement between the ML model prediction and experimental test and DEM simulation results indicates that the ML–based modelling approach is capable of capturing accurately the effects of initial void ratio on the constitutive behaviour of idealised granular materials, bypassing the need to incorporate the complex micromechanics underlying the macroscopic mechanical behaviour of granular materials. Lastly, a detailed comparison between the used MLP model and long short-term memory (LSTM) model was made from the perspective of technical algorithm, prediction accuracy, and computational efficiency.

Little research can be found in relation to the stability of anisotropic and heterogenous soils in three dimensions. In this paper, we propose a study on the three-dimensional (3D) undrained slopes in anisotropic and heterogenous clay using advanced upper and lower bounds finite element limit analysis (FELA). The obtained stability solutions are normalized, and presented by a stability number that is a function of three geometrical ratios and two material ratios, i.e. depth ratio, length ratio, slope angle, shear strength gradient ratio and anisotropic strength ratio. Numerical results are compared with experimental data in the literature, and charts are presented to cover a wide range of design parameters. Using the multivariate adaptive regression splines (MARS) analysis, the respective influence and sensitivity of each design parameter on the stability number and the failure mechanism are investigated. An empirical equation is also developed to effectively estimate the stability number.