Large quantities of gold tailings are generated and stockpiled worldwide, creating significant environmental and land-use challenges. However, the application of gold tailings powder (GTP) in high-performance engineered cementitious composites (HP-ECC) remains limited, and its effects on mechanical performance, fiber–matrix interfacial behavior, and microstructural evolution require further investigation. This study investigated the feasibility of using activated gold tailings powder (GTP) as a partial cement replacement (0–30%) in HP-ECC. The effects of GTP content on flowability, setting time, mechanical properties, pore structure, and fiber–matrix interfacial behavior were systematically evaluated. GTP reduced the flowability of fresh mixtures, but an appropriate replacement level improved the overall performance of HP-ECC. The optimum performance was achieved at a replacement level of 20%, where the compressive strength, flexural strength, and ultimate tensile strength reached 97.4 MPa, 19.4 MPa, and 8.16 MPa, respectively. The tensile strain capacity ranged from 3.93% to 5.55%, and all mixtures exhibited strain-hardening behavior with multiple cracking. Single-fiber pullout tests showed that the interfacial frictional bond stress of the GTP-20 mixture reached 1.79 MPa. Microstructural analysis suggested that the filler effect and potential pozzolanic activity of GTP contributed to pore structure refinement, porosity reduction, and enhanced fiber–matrix interfacial properties. These findings demonstrate that activated GTP is a viable supplementary cementitious material for HP-ECC and a promising option for the value-added utilization of gold tailings.
Accurate backfill grouting plays a critical role in treating soil karst cave foundations. However, the inherent concealment of underground construction complicates the assessment of actual grouting and filling conditions. Digital construction technology facilitates the dynamic monitoring of grouting parameters, including slurry flow, pressure, and water-cement ratio, providing essential field data for analyzing grout diffusion within soil karst caves. By leveraging these digital monitoring data, a computational-fluid-dynamics-based numerical simulation was conducted to investigate grout flow, diffusion, and filling patterns during the grouting process. A foundational three-dimensional geological model for soil karst cave foundation treatment was developed by integrating subsurface geology, dynamic grouting hole data, and topography. By coupling this geological model with real-time grouting monitoring data and employing the Bingham constitutive model for cement grout, an integrated digital simulation method for the coupling of two stages of low-grade concrete pouring and sleeve-valve pipe grouting under digital construction was proposed. This methodology utilizes the finite difference method, true volume of fluid technique, and porous media theory. A sensitivity analysis was also performed to examine the influence of grouting pressure and water-cement ratio on the grout diffusion time and injection volume during filling. The findings demonstrate strong agreement between the simulated values of grout flow, filling, and diffusion within the karst cave and on-site measurements (the errors in the two-stage infusion volume were 6.4% and 5.0%, respectively). This integrated approach provides an effective means of optimizing grouting parameters and enables comprehensive quality control throughout the soil karst cave foundation grouting process.
The loads applied on walls of a foundation pit are frequently asymmetric due to complicated construction environment and conditions. Relationship between earth pressure and lateral deformation is quite complex, resulting in difficulties in proposing theoretical solution to lateral deformation of diaphragm wall under asymmetric loads. In this study, the deformation of the internal supports and the value of earth pressure are coupled with the lateral deformation. A generic third-order plate theory is employed to construct the theoretical model for calculating the lateral deformation of the two diaphragm walls of a foundation pit under asymmetric loads. The governing equation is obtained based on the principle of minimum potential energy. An iterative method is proposed in order to consider both the active and passive earth pressures under different lateral deformation. The analytical solution for the lateral deformation is obtained by utilizing the Pb-2 Rayleigh-Ritz method.The accuracy of the theoretical model is verified by comparing it with the monitoring data of the “New World” pit in Hangzhou, China, with the mean absolute error (MAE) of 1.23∼1.47 mm and the coefficient of determination (R²) of 0.943∼0.958, which is 49.3%∼85.8% higher than the classical plate theory in prediction accuracy.The research results can provide a new theoretical method for the deformation calculation and optimization design of diaphragm walls under asymmetric loads in deep foundation pit engineering.
Personal Mobility Devices (PMDs), also known as micromobility devices, have expanded rapidly in Korea, accompanied by a sharp increase in accidents and severe injuries. Using six years of nationwide PMD accident data, this study applies a decision tree approach to identify key factors and their non-linear interactions across human, environmental, and accident-related variables. The results show that accident type is the most influential factor, followed by rider age and gross negligence. PMD-only accidents, although the least frequent, exhibit the highest fatality risk, particularly among older riders and unlicensed young riders. PMD-to-vehicle accidents are the most common, with higher fatality risks observed in head-on crashes and turning maneuvers involving older riders, while younger riders face greater risks at large intersections. PMD-to-pedestrian accidents show the lowest fatality rate but greater severity when gross negligence is involved. These findings highlight the need for targeted training, infrastructure improvements, and policy measures to enhance PMD safety.
Conventional ecological slope-protection substrates rely heavily on planting soil, limiting the large-scale utilization of industrial solid waste. This study developed a solid-waste-based substrate using fly ash as the principal substitute for peat soil and a composite binder of flue gas desulfurization gypsum (FGDG) and cement. An L16 orthogonal design and the entropy weight method were used to evaluate the effects of mixture proportions on physicochemical, vegetation-related, and mechanical performance. The optimal formulation contained 30% fly ash, 12% cement, 15% FGDG, and 43% peat soil, with 2% water-retaining agent and 2% biochar. The total solid-waste content exceeded 50% while maintaining satisfactory vegetation growth and mechanical stability. Simulated rainfall tests showed that runoff increased with rainfall intensity and slope gradient, whereas infiltration was limited by the substrate infiltration-capacity threshold. Under identical rainfall and slope conditions, soil loss from the solid-waste substrate was approximately 60% lower than that from the peat-soil substrate. Regional assessment using the Chinese Soil Loss Equation classified 30°–50° slopes in Dalian as having very slight or slight erosion. The developed substrate can therefore reduce planting-soil consumption while providing adequate erosion resistance for ecological slope protection.
Steel tube girder bridges are widely adopted for pedestrian crossings owing to their relatively short construction time and superior aesthetics. However, conventional circular steel tube girders suffer from poor load distribution at the girder-deck interface, inadequate deck plate rigidity, and critical quality problems associated with on-site welding of longitudinal connections. To address these limitations, this paper proposes a corrugated steel tube girder with an innovative prefabricated interlocking connection system and validates its performance through three sequential phases: (1) design concept and connection details, (2) analytical verification by finite element analysis, and (3) experimental validation by static load testing. Composite action between the corrugated steel tube and the concrete deck plate shifts the neutral axis upward, reducing compressive stress in the steel tube to 44% of the conventional value and achieving an approximately 48% reduction in steel consumption. On-site connection time is substantially reduced by eliminating continuous field welding. Finite element analysis using MIDAS Civil confirmed that all member stresses remained within allowable limits under six load combinations per KDS 24 14 30. Maximum girder deflection of 4.795 mm was well within the 25 mm allowable limit, and the fundamental frequency of 2.75 Hz exceeded the resonance-critical range. Static load testing of a reduced-scale specimen (1:2 cross-sectional scale) reached a maximum applied load of 236.3 kN, corresponding to 1.18 times the prescribed ultimate load of 200 kN, with a steel strain safety factor of at least 2.08 and a concrete deck safety factor of at least 2.95. No joint opening or fracture was detected by the installed displacement measurements up to the maximum applied load.
During the construction within active fault zones in tunnels, challenges like collapse and deformation may arise. This study is grounded in field experiments, aiming to analyze the stress and deformation characteristics of Dongmachang No.1 tunnel (DMT-1). Leveraging machine learning techniques, the research predicts the tunnel's settlement values over the following two years. Furthermore, numerical simulations assess the support reliability of ultra-high-performance concrete (UHPC) lining, validated through on-site monitoring. Findings reveal that significant geostress and the presence of active fault zones intersecting the deeply buried DMT-1 lead to deformation manifestations such as primary support deformation, arch cracking and uplift, and secondary lining cracking. The proposed UHPC lining demonstrates effective tunnel reinforcement, playing a pivotal role in the sustainable development of tunnel infrastructures.
Public parking demand in Korean cities has been inadequately modeled because prior studies rely on administrative-district aggregation and zoning- or floor-area-based land-use variables decoupled from actual parking-generating activity. This study introduces 165 parcel-level building-use codes from the Korean building-property tax ledger, aggregated to 137 small blocks in Suseong-gu, Daegu, for the years 2020 and 2023. Factor analysis with Ward's hierarchical clustering identifies three interpretable clusters, and market-segmentation regressions are estimated separately for daytime and nighttime demand. Segmentation improves all reproducibility metrics (MAE, RMSE, MAPE, χ², r) relative to the pooled model across all (year × time) combinations, and a χ² homogeneity test rejects model equivalence at p < 0.001. Medium-density residential floor area is the strongest predictor of nighttime demand in both aggregate and segmentation models. Food-and-wedding floor area leads daytime demand in the aggregate model, while medical and commercial uses are significant daytime predictors within cluster models. Proximity to urban-rail stations is positively associated with daytime demand in the mixed-residential cluster, consistent with a co-agglomeration mechanism obscured by the aggregate model. The findings support time-of-day-differentiated policy: nighttime capacity expansion near medium-density residential zones and daytime expansion near activity-intensive uses. The pipeline is reproducible wherever comparable building-use records exist.
This study compiled 5,133 records for machine-learning model development, consisting of 3,686 Korean field records and 1,447 curated records from the international literature. A separate temperature-dependent dataset contained 726 age–strength measurements from 27 distinct mixture designs. The proposed hybrid framework combines a physically constrained ultimate-strength model, machine learning, and the time- and temperature-dependent apparent activation energy (T–TaAE) formulation. The T–TaAE parameters were jointly calibrated across all available curing temperatures for each mixture design. Grouped calibration achieved R2 = 0.955 and RMSE = 3.32 MPa, while leave-one-temperature-out validation achieved R2 = 0.922 and RMSE = 4.36 MPa. The model is implemented in a Streamlit-based research platform that reports predicted strength and a model-derived 90% prediction interval. Its present use is limited to decision support within the validated mixture, age, and nominal temperature ranges; it is not presented as a safety-validated substitute for field strength testing.
Urban areas contribute disproportionately to global carbon emissions, highlighting the need for multifunctional nature-based solutions. Constructed wetlands (CWs) are widely implemented for stormwater management and ecosystem service provision, yet their role in soil organic carbon (SOC) regulation remains insufficiently understood. This study evaluated the relationships among soil physicochemical properties, microbial community composition, and SOC storage in urban stormwater CWs. Soils from three CW systems were compared with an adjacent natural landscape using standard physicochemical analyses and 16S rRNA gene sequencing. SOC accumulation was highest in the mature surface flow constructed wetland (SFCW), which may be associated with vegetation development and sustained stormwater inputs. Proteobacteria dominated across sites, and SOC was positively associated with specific microbial groups, while soil bulk density and pH were related to microbial community patterns. Although microbial diversity was greater in the natural landscape, constructed systems showed distinct soil–microbial patterns that may be linked to SOC retention. These findings suggest that mature urban CWs may have potential to support soil carbon storage, but the observed relationships should be interpreted as associative rather than causal. However, net climate regulation or carbon balance could not be evaluated from SOC data alone without greenhouse gas flux measurements, particularly CH₄ and N₂O.
Elevated silos face seismic damage risks due to their raised mass centers. To mitigate production interruptions and safety accidents caused by collapse failures, this study employs a top isolation system (NRB-UDs) for retrofitting elevated silos. The system combines natural rubber bearings (NRB) with bidirectional U-shaped dampers (UDs), arranged circumferentially between the top of the support frame and the silo stiffening ring. First, a finite element model of the elevated silo is established, using 3D plastic solid elements to simulate the stored material. Displacement loading tests up to 100% shear strain are conducted on the NRB-UDs system to calibrate an equivalent bilinear model; the isolation system is then simulated using parallel Connector elements and Gap elements. Finally, incremental dynamic analysis is performed on both the original and retrofitted structures. Seismic demand parameters are fitted based on the maximum inter-story drift ratio and spectral acceleration Sa(T1, 5%), and fragility curves are developed. The results show that the seismic failure of elevated silos is due to premature yielding of the support system and the progressive formation of plastic regions, leading to soft-story behavior. The NRB-UDs system exhibits stable energy dissipation and an equivalent damping ratio. After retrofitting, under a seismic intensity with a 10% probability of exceedance in 50 years, the probability of the structure reaching severe damage is reduced from 97% to 38%. The collapse margin ratio increases from 0.85 to 8.10, with support members remaining elastic, effectively eliminating the soft-story mechanism. The top isolation scheme provides a cost-effective retrofitting strategy for elevated silos.
End-suspended wall support systems require high stability of the underlying rock wall section, especially in foundation excavations without rock shoulders, yet existing specifications provide limited guidance on stability evaluation. Based on the deep foundation excavation of Lingnan Square Station on Guangzhou Rail Transit Line 12, this study applies the upper-bound limit analysis method to evaluate the stability of rock wall sections supported by end-suspended walls without rock shoulders. A strength reserve safety factor is introduced to derive analytical formulas for the safety factor and required supporting force, and the assumed failure mechanisms are qualitatively assessed using numerical displacement contours and convergence behavior. The results indicate that stability is governed by two dominant failure mechanisms: a combined rotation and translation mode and a combined translation, rotation and translation mode. The calculated safety factor ranges from 0.12 to 2.72 and from 0.36 to 2.03 under the two modes, respectively. Parametric analysis shows that increasing the uniaxial compressive strength of intact rock from 4 to 20 MPa increases the safety factor by more than 50%. Grey relational analysis shows that the shear strength of the underlying rock primarily controls the combined rotation and translation mode, while the depth of the rock-retaining section dominates the combined translation, rotation and translation mode. The proposed approach provides a practical theoretical basis for the design and stability evaluation of end-suspended wall-supported foundation excavations in soil-rock composite strata.
Rock mass stability in underground excavations is strongly influenced by the geometry of pre-existing fractures. This study investigates the mechanical response and fracture evolution of granite specimens containing two unequal-length prefabricated fractures with different secondary fracture dip angles (SFAs) under uniaxial compression. Digital image correlation (DIC) and acoustic emission (AE) monitoring were used to characterize the coupled evolution of surface deformation and internal damage. The results show that all specimens exhibit a typical four-stage stress–strain response. The peak compressive strength varies non-monotonically with SFA, increasing by 35.59% from 30° to 60°, decreasing by 34.55% from 60° to 120°, recovering by 32.73% at 135°, and subsequently decreasing by 25.70% at 150° DIC results indicate that strain localization initiates mainly at the fracture tips and within the rock bridge, while the dominant failure mode changes from predominantly tensile cracking to mixed tensile–shear cracking depending on fracture geometry. AE activity exhibits a three-stage evolution pattern, and both the onset of rapid AE activity and the b-value evolution show clear angle dependence. These findings highlight the critical role of secondary-fracture orientation in crack interaction, damage evolution, and macroscopic failure, providing useful insights into the stability assessment of jointed rock masses.
To address the problem of low mechanical excavation efficiency for cutter in extremely hard rock, the rock-breaking influence law of microwave-assisted TBM cutter mode is investigated. Rock-breaking experiments using microwave pretreatment and small-scale cutters were conducted. The fragment size distribution, cutting forces, and energy efficiency for microwave-assisted cutter were analyzed. And the influence of microwave parameters on the rock-breaking performance has been researched. Results show that under an equivalent energy input of 900 kJ, the 3kW-5 min group achieved a 41.94% increase in rock-breaking volume and a 36.34% reduction in specific energy compared to the 1kW-15 min group. Evidently, the high-power, short-time scheme significantly outperforms the low-power, long-time alternative. Furthermore, the total specific energy index identified the 3kW-5 min combination as the optimal energy efficiency window, with a total energy consumption of 135.3 MJ/m³. And the normal and rolling forces reduced by 13.66% and 15.38%, respectively, relative to the untreated condition. The study provides a quantitative scientific basis for the parameter optimization of microwave-assisted cutter breaking rock system for TBM.
This study developed a machine learning–based predictive framework to preemptively classify four major types of bridge deck pavement damage—cracking, potholes, alligator cracking, and pavement wear—and elucidated their physical deterioration mechanisms using SHAP (SHapley Additive exPlanations) analysis. A comprehensive dataset was constructed by integrating 12,972 pavement inspection records collected from 3,162 bridges in South Korea between 2004 and 2024 from the Bridge Management System (BMS), together with traffic data from the Traffic Monitoring System (TMS) and climatic variables from the Korea Meteorological Administration (KMA). Three machine learning algorithms—XGBoost, Random Forest, and Logistic Regression—were compared under three class imbalance strategies (No-Resampling, SMOTE, and ADASYN), with hyperparameters optimized using Optuna and model robustness validated through 10-fold cross-validation and a hold-out test dataset. XGBoost consistently outperformed the other algorithms across all damage types. For cracking and potholes, which exhibit distinct morphological boundaries, the original dataset (No-Resampling) yielded the highest F1-Scores of 0.8783 and 0.8227, respectively. For progressive damages with ambiguous boundaries—alligator cracking and pavement wear—SMOTE oversampling achieved the best performance, with F1-Scores of 0.8150 and 0.7644, respectively. All models achieved high precision (0.89–0.98), minimizing false-positive detections that could lead to unnecessary maintenance expenditures. SHAP analysis quantitatively identified the dominant deterioration drivers for each damage type, including thermal stress and heavy truck traffic (AADTT) for cracking, moisture infiltration and freeze–thaw cycles for potholes, cumulative fatigue under sustained loading for alligator cracking, and time-dependent UV aging combined with winter maintenance operations for pavement wear. These findings provide data-driven evidence to support the transition from reactive to preventive bridge infrastructure asset management.
This study develops an explicit closed-form equation for predicting the natural frequencies of multi-stiffened square plates with all four edges fixed. A comprehensive numerical database is generated by using finite element analysis systematically varying plate dimensions, stiffener configurations, and stiffener geometrical properties. Genetic Expression Programming (GEP) is employed to derive an explicit and user-friendly empirical formula, while several supervised machine learning algorithms as ridge regression, artificial neural networks, random forest, and gradient boosting are utilized for performance comparison. The proposed models are validated against finite element results using statistical indicators, demonstrating that the GEP-based formulation achieves high accuracy with low prediction errors. Parametric and sensitivity analyses are also conducted to investigate the influence of key design variables on the natural frequencies. The results indicate that the developed empirical expressions provide an efficient and practical tool for preliminary design and dynamic assessment of stiffened plate structures.
This paper proposes a physics-informed neural operator framework, termed physics-informed neural operator-vehicle-bridge interaction (PINO-VBI), for efficiently predicting the dynamic responses of vehicle-bridge interaction (VBI) systems. The numerical validation focuses on controlled low-speed VBI scenarios, in which the vehicle speed ranges from 4 to 8 m/s. The model integrates a branched Fourier neural operator (BFNO) with physics-informed loss terms to accommodate the frequency heterogeneity among vehicle subsystems and incorporate governing differential equations during training. To address the disparity in response characteristics across subsystems, a dual-branch BFNO architecture is developed, allowing flexible allocation of Fourier modes and hyperparameters for different frequency components. The network is trained using synthetic datasets generated from VBI simulations, incorporating variations in vehicle speed, system parameters, and random road excitations. A comprehensive hyperparameter study is conducted to assess the sensitivity of prediction performance to model architecture. Furthermore, the impact of embedding different types of physical constraints, including equation residuals, velocity constraints, and acceleration constraints, is systematically analyzed. Comparative experiments demonstrate that the proposed PINO-VBI outperforms standard FNO-based PINO models, particularly in accurately capturing velocity and acceleration responses critical to ride comfort and structural safety assessment. This study highlights the advantages of combining physics-informed learning with frequency-aware neural operators in modeling VBI systems, and provides a methodological basis for efficient surrogate simulation and future digital twin applications in bridge health monitoring and dynamic safety evaluation.
Driven by the digital transformation of the construction industry, multimodal vision-language models have been widely applied to construction scene perception. However, mainstream methods are constrained by poor domain semantic adaptation, low cross-modal matching accuracy, and high manual annotation costs. This paper proposes a BLIP-2 multimodal scene understanding method with a self-accumulated semantic mechanism. A semantic progressive domain fine-tuning strategy is adopted based on indoor construction image-text datasets to improve the model’s descriptive capability for construction-specific objects and behaviors. Experimental results on an indoor renovation corridor show that the optimized model achieves BLEU, ROUGE-L and METEOR scores of 0.14, 0.44 and 0.35, respectively, with a key scene element recall rate of 0.89, substantially outperforming the original BLIP-2 model. Validated for narrow indoor construction scenarios, the proposed lightweight model serves as a front-end semantic perception module for digital twin systems. It supports fine-grained object recognition, scene semantic abstraction, real-time environmental perception, construction behavior analysis and risk early warning in intelligent construction. The framework delivers stable and quasi-real-time image parsing in the test scenario, while its robustness and cross-scenario scalability require further experimental verification. Compared with traditional manual-dependent and static visual detection schemes, this method reduces annotation biases and provides an innovative approach for intelligent construction site monitoring.
A single mixed passenger and freight railway adopts continuous welded rail (CWR) on a super-large-span deck arch bridge with a π-shaped rigid frame, and the main span length is 570m. To support the design of CWR on the deck arch bridge, the influence of bridge parameters on the forces and deformation of CWR on the deck arch bridge were investigated. The results show that the beam-span arrangement on the arch significantly affects the longitudinal forces in the CWR on the deck arch bridge, and a reasonable beam-span arrangement can effectively reduce the longitudinal force. When the longitudinal horizontal stiffness of columns increase from 94.4kN/cm to 10000kN/cm, the rail expansion force, rail bending force, rail braking force and rail broken gap decreased by 1.43%, 1.24%, 1.22% and 1.79% respectively. The arch rib stiffness has a much stronger effect on the rail bending force than on the rail expansion force and rail braking force, and the reduction of rail bending force could reach up to 68.53%. It is recommended that one rail expansion joints be installed on each side of the arch rib for the CWR on the super-large-span deck arch bridge.
The structural integrity and load-bearing capacity of reinforced concrete structures are fundamentally governed by the bond interaction between steel bars and concrete. This study focuses on the bond stress-slip behavior of HRB650E seismic steel bars, which possess distinctive mechanical properties including elevated yield strength and enhanced ductility compared to conventional steel bars. These characteristics significantly influence its interfacial behavior with concrete, necessitating specialized investigation. To systematically evaluate the bond characteristics of HRB650E steel bars, an extensive experimental program was implemented involving 96 large-scale specimens organized into 24 distinct test groups. The experiment was designed to examine the effects of five critical parameters: concrete strength, concrete cover thickness, anchorage length, bar diameter and stirrup ratio. The results revealed that ultimate bond strength demonstrates positive correlations with concrete strength, cover thickness, and stirrup ratio, while exhibiting an inverse relationship with anchorage length. Regarding deformation characteristics, ultimate bond slip was found to increase with both cover thickness and stirrup ratio.Based on quantitative analysis of experimental results, this research establishes several contributions: (1) a design equation for determining the basic anchorage length of HRB650E steel bars; (2) predictive equations for estimating ultimate bond strength and bond slip; and (3) a refined bond-slip constitutive model that effectively captures the behavioral trends observed in testing. The proposed model shows strong correlation with experimental measurements. These developments provide valuable tools for optimizing the design of concrete structures incorporating HRB650E high-strength seismic steel bars, potentially leading to more efficient and reliable structural configurations.
