Bulletin of Engineering Geology and the Environment最新文献

Slope failures are the leading cause of significant casualties and financial losses, despite technological advancements. Digital Twin has become an essential tool for real-time monitoring and decision-making, particularly for the complex problems related to slope stability. Data-driven algorithms have proven to be efficient surrogate models for Digital Twins as an alternative to the existing limit equilibrium and finite element analysis. In this study, ensemble learning algorithms are employed to predict the factor of safety (FoS) of the slopes, and Explainable Artificial Intelligence (XAI) methods are used to improve the model interpretability. The findings of the study are aimed at identifying key features important for effective monitoring of slopes in the field, showing that XAI findings align well with established physical principles of slope. A comprehensive field database containing 497 unique slope cases consisting of both manmade and natural slopes with circular failure conditions is utilised in this study for model development and verification. Soil unit weight (γ), soil cohesion (c), internal friction angle (ϕ), slope angle (β), slope height (H), and pore pressure ratio (r_u) are taken as primary input parameters. Notably, features H, c, and β are consistently ranked as the most important features across models, highlighting their crucial function in slope stability. A structured approach is used to rank the models, showing that combining ensemble learning and XAI within Digital Twin framework can lead to safer, data-driven decisions for slope stability issues.

Plastic films randomly distributed in landfills can enhance soil strength through friction and interlocking. However, the evolution of their orientation distribution under varying loads and the methods to quantify such effects remain poorly understood. This study employed X-ray computed tomography (CT) to characterize the spatial distribution of plastic film fibers under confining pressures of 50–150 kPa. A reinforcement calculation method was developed that explicitly accounts for fiber shape, tensile properties, and measured distribution characteristics. Results showed that the gamma distribution effectively described the bimodal and skewed angle distributions of fibers. Empirical relationships between distribution parameters ((a), (b)) and applied load ((P)) were established (R² > 0.90), and a functional expression relating orientation angle (theta) to (P) was proposed. The framework predicted strength parameters for 0.2% mass content of films as (c)= 19.08 kPa and (varphi)= 34°, closely matching experimental values ((c)= 18.11 kPa, (varphi)= 34.2°), with relative errors of 5.4% and 0.58%, respectively. These results demonstrate the feasibility of quantifying the reinforcement contribution of plastic film fibers in landfill soils under the tested conditions.

Graphical Abstract

Excavation in fractured rock masses often induces stress redistribution and progressive fracturing, which can jeopardize the stability of engineering foundations. This study develops a three-dimensional Discrete Element Method (DEM) model using the Flat-Joint contact law in PFC3D to simulate excavation-induced crack initiation and propagation within the sandstone formations of the Madar anticline, southwest Iran. Micro-mechanical parameters were calibrated through iterative back-analysis of uniaxial compression (UCS = 39.59 MPa) and Brazilian tensile strength (BTS = 4.82 MPa) tests, achieving less than 7% deviation from laboratory results. Staged excavation was simulated in three sequential 2.5 m steps, enabling realistic stress release and fracture coalescence. The results show that tensile cracking dominated during the first stage, accompanied by a surface displacement of about 3.6 cm, while progressive shear coalescence in the second stage increased displacement up to 8 cm; the final stage exhibited partial stress relaxation with displacements below 1.5 cm. The simulated fracture propagation followed the orientation of the predefined field-based discontinuities, showing consistent fracture reactivation rather than an independent prediction of fracture orientation. Localized tensile stresses exceeding 3.2 MPa occurred near the excavation front, explaining early microcrack nucleation. The results demonstrate qualitative agreement between the DEM–FJM simulations and field observations based on fracture distribution, displacement magnitude and crack evolution.

Avalanches are complex natural hazards whose dynamics and destructive capacity vary widely with snowpack properties, terrain, and climate. This study investigates the formation characteristics, mechanisms, and dynamic responses of different avalanche types in the northern Himalayas of Tibet. Field observations, including video monitoring and snow property measurements, were conducted to characterize avalanche development and obtain key physical parameters. Based on measured snow density and cohesion, numerical simulations using the Rapid Mass Movement Simulation (RAMMS) model were performed to compare the dynamic characteristics of powder, flow, and slab avalanches. The effects of erosion entrainment and terrain conditions on avalanche dynamics and impact pressure were further evaluated. Results indicate that powder avalanches are characterized by high mobility, short movement duration, and relatively low impact pressure due to weak basal interaction and limited erosion capacity. Slab avalanches exhibit stronger destructive potential, with higher impact pressure and more pronounced erosion-induced volume amplification resulting from their dense structure. Flow avalanches show intermediate dynamic characteristics between powder and slab avalanches. The simulations further reveal that accounting for erosion processes alters avalanche motion and affects predicted velocity and impact pressure. Moreover, terrain characteristics strongly regulate avalanche impact pressure, with steep, concave, and smooth terrains promoting higher impact forces, whereas gentle, convex, and rough terrains reduce impact intensity. This study improves understanding of the dynamic mechanisms and destructive effects across avalanche types. It provides a scientific basis for avalanche risk assessment and the optimization of mitigation strategies in the northern Himalayas region of Tibet.

To better understand the behavior of liquefiable sand during the process of liquefaction, strain-controlled cyclic triaxial tests were performed on saturated poorly graded fine sand. Specimens with relative densities (({D}_{r})) from loose to medium-dense states were tested under various effective confining pressures (({sigma}_{o}^{^{prime}})) corresponding to an overburden depth < ~ 10 m. Results show that the excess pore water pressure (EPWP) buildup versus loading cycles ((N)) depends on ({D}_{r}), ({sigma}_{o}^{^{prime}}), and the cyclic shear strain amplitude (({gamma}_{c})). Moreover, a certain threshold of ({gamma}_{c}) is required for liquefaction-level EPWP buildup within realistic earthquake loading cycles; this threshold also depends on ({D}_{r}) and ({sigma}_{o}^{^{prime}}). Using the modified two-step parameter calibration procedure, these relationships are satisfactorily approximated by the Vucetic–Dobry EPWP generation model. The shear modulus decreases with increasing ({gamma}_{c}) and (N), reflecting its link to EPWP buildup; accordingly, normalized relationships between shear modulus degradation and EPWP buildup are established. This degradation accelerates with increasing EPWP buildup and becomes substantial near liquefaction, aligning with existing laboratory cyclic loading tests on similar materials. Notably, these normalized relationships exhibit little dependency on ({D}_{r}) and ({sigma}_{o}^{^{prime}}). However, such soil modulus degradation can differ significantly from the reduction in soil resistances to foundations observed in foundation loading tests. Hence, varying degrees of liquefaction-induced soil degradation across different engineering applications should be incorporated into seismic design.

Due to arid climate and short-duration intense rainfall in summer, vast number of earthen cultural heritage structures in northwestern China are often subjected to wetting-drying (W-D) cycles. Earthen plaster, primarily made of clay mixed with additives (sands, fibers), are often used as the external skin of many historical buildings. These earthen plasters frequently develop numerous cracks under W-D cycles, which is detrimental to cultural heritage conservation. To clarify the influence of W-D cycles on crack development in earthen plaster, this study investigated the effects of plaster material composition (coarse particles, two types of fibers, and three types of binders) and three W-D cycles on the formation and development of cracks through W-D cycles test and crack image analysis. The results show that crack development in earthen plasters is influenced by both material composition and the number of W-D cycles. The addition of fibers (wheat straw and hemp fiber) effectively inhibits crack formation and promotes crack healing during wetting. Although the initial crack width is larger after adding natural hydraulic lime to earthen plaster, it still reduces the crack ratio of the plaster and gives it better resistance to W-D cycles. These findings hold significant implications for material selection in earthen plaster and conservation strategies in the restoration of earthen plasters in cultural heritage structures, particularly in regions subjected to extreme conditions and intense W-D cycling.

Efficient utilization of mine geothermal energy is an important pathway for the green development of coal mines and the clean energy transition of mining areas. In this study, the heat-accumulation mechanism and development potential of hydrothermal geothermal resources in the Pingdingshan No. 8 mine were investigated using hydrogeochemical, isotopic, borehole-temperature, and numerical-simulation data. A geothermal well-cluster development model was proposed. The results show that the Cambrian limestone geothermal water is mainly characterized by Na-Ca-HCO₃-SO₄ and Na-HCO₃-SO₄ hydrochemical types, with an average reservoir temperature of 70.43 ℃. Isotopic data indicate that meteoric water, especially rainfall during wet periods, is the primary recharge source. The average geothermal heat flow is 70.99 mW/m², and the average borehole-derived geothermal gradient is 33.6 ℃/km. Geothermal accumulation is mainly controlled by high thermal conductivity in uplifted structural zones, compound fold structures, and convective heat transfer along deep-seated faults. The base-case recoverable geothermal resource is estimated to be 5.47 × 10¹⁷J, which can support heating for approximately 100,000 m² of building area for 62.5 years under the simulated 200 m³/h pumping–recharge scenario. Compared with the single-well case, the well-cluster system delayed thermal breakthrough by 23 years under the specific simulated conditions. Replacing coal-fired heating with geothermal utilization can reduce CO₂ emissions by approximately 9,970 t annually. This study provides a scalable mine geothermal utilization framework for supporting low-carbon energy transition in mining areas.

The typical practice in dam construction is to complete the grout curtain before reservoir impoundment; however, in some cases, grouting may continue during or after impoundment. This study compares the statistical distribution of grout takes before and during reservoir impoundment at the Seymareh Dam using two geologically comparable grout curtain galleries (GL3 and GL4). These galleries were selected because they intersect similar lithological units of the middle and lower Asmari Formation and exhibit comparable structural characteristics, while differing primarily in the hydraulic conditions prevailing during grouting. Approximately 175,000 m of grout curtain records from the 180-m-high Seymareh Dam were analyzed using Ewert's (2008) unpublished methodology. Unlike previous studies that focused exclusively on pre-impoundment conditions, this research provides a systematic field-scale comparison between grouting performed before and during reservoir impoundment, enabling the identification of grout washout mechanisms in large dam foundations. Our analysis reveals significant differences in grout-take distributions between the two hydraulic conditions. Grout take increased by up to five times in boreholes affected by reservoir impoundment, indicating a strong hydraulic control on grout consumption under rising reservoir head conditions. Under normal hydraulic conditions, the observed grout-take distributions agree well with Ewert's conceptual patterns for wide and extended joints (Type b), fissure-karst conditions (Type c), and small joints with incidental cavities (Type e), reflecting the engineering geological conditions encountered at the Seymareh Dam site. In contrast, grout-take distributions obtained during reservoir impoundment deviate systematically from these expected patterns because of the additional influence of transient hydraulic loading. These results indicate that hydraulic loading acts as an additional controlling factor superimposed on geological conditions, a factor that is not explicitly incorporated into existing classification frameworks. These findings introduce a practical criterion for detecting grout washout based on deviations from Ewert's statistical distribution patterns. Rather than replacing Ewert's conceptual framework, the present study extends its practical applicability by demonstrating its limitations under transient hydraulic loading during reservoir impoundment. The results further indicate that, in addition to geological conditions, reservoir impoundment and grouting sequence significantly influence grout-take distributions.

Internally unstable soils, consisting of coarse and fine particles where the coarse skeleton fails to retain fines, are highly prone to internal erosion under seepage. Such erosion degrades both hydraulic and mechanical properties, threatening the long-term stability of natural and engineered soil structures. In practical situations, soils are frequently subjected to fluctuating hydraulic conditions. To investigate the effects of hydraulic fluctuations on internal erosion and post-erosion stress-strain behavior, an improved triaxial erosion apparatus was employed to conduct internal erosion and post-erosion triaxial compression tests under different controlled hydraulic loading patterns. Test results demonstrated that hydraulic gradient fluctuations significantly exacerbate internal erosion, with increased fine particle loss under higher fluctuation amplitudes and frequencies. It was found that the variations in hydraulic conductivity and volume deformation induced by fine particle loss can be correlated with the fines eroded ratio and confining pressure, independent of the specific hydraulic loading paths and patterns. Substantial alterations in post-erosion stress-strain behavior were observed, with the shear dilatancy and peak friction angle decreasing with increasing fines eroded ratio. The reduction in the peak friction angle can also be well related to the fines eroded ratio. These findings improved the understanding of erosion-induced deformation in internally unstable soils under fluctuating hydraulic gradients.