Search Results (367)

The fractal characteristics of coal pore–fracture networks and their evolution under compression are essential for predicting rock mass failure and fluid transport. This study combines micro-CT scanning with fractal theory and seepage mechanics to investigate the structural evolution of coal under uniaxial compression and its impact on fluid transport. CT scans were performed at four characteristic stages (initial, elastic, plastic, and failure) to reconstruct three-dimensional fracture networks. Quantitative analysis reveals that fracture porosity increases sequentially from 0.44% to 5.01%, with the failure stage reaching 11.4 times the initial value. Fracture length and aperture distributions follow power-law scaling, and their fractal dimensions exhibit distinct evolution patterns: length dimension increases from 2.43 to a peak of 2.56 in the plastic stage and then drops to 2.47 at failure, while aperture dimension decreases from 2.29 to a trough of 2.12 before rebounding to 2.26. These patterns reflect a dynamic adjustment of network complexity, transitioning from primary fractures to micro-fracture dominance and finally to main fracture coalescence. Based on the Knudsen number, three diffusion regimes of Fick, transition and Knudsen are identified. A fractal permeability model is developed by idealizing the pore space as tortuous capillaries, showing that permeability scales with the fourth power of the maximum pore diameter and is positively influenced by the fractal dimension and the number of large pores. Furthermore, a coupled seepage–stress model is derived, incorporating pressure transmission, shear transmission, and crack opening coefficients. The damage variable is expressed as a function of stress level and fractal dimension. These findings provide theoretical support for predicting gas transport and failure behavior in coal under coupled hydro-mechanical conditions. Full article

Wittichenite (Cu₃BiS₃) was synthesized by mechanical alloying (MA) followed by hot pressing (HP), and its phase evolution, thermal stability, charge transport behavior, and thermoelectric performance were systematically examined. X-ray diffraction analysis of the MA powders revealed broadened diffraction peaks, indicating reduced crystallinity and refined crystallite size. After HP consolidation, a well-defined single-phase orthorhombic wittichenite structure was obtained. These results demonstrate that the mechanically induced solid-state synthesis was effectively initiated during MA and subsequently completed through crystallization, defect relaxation, and densification during HP. The MA–HP processed specimens exhibited high relative densities of 94–98% of the theoretical value and a homogeneous microstructure without detectable compositional segregation or grain-boundary enrichment, confirming the formation of a structurally and chemically stable single-phase bulk material. Thermal analysis identified a reversible polymorphic phase transition from P2₁2₁2₁ to Pnma at low temperature, followed by structural relaxation and the onset of partial decomposition at higher temperatures, indicating that Cu₃BiS₃ retains structural integrity below 700 K, which defines the relevant operating window for thermoelectric evaluation. The samples exhibited p-type semiconducting behavior, with electrical conductivity increasing with temperature due to thermally activated hole transport and showing an additional enhancement across the structural transition region. The Seebeck coefficient remained positive over the entire temperature range and decreased gradually with increasing temperature, consistent with semiconductor transport characteristics. The thermal conductivity remained low at 0.30–0.38 W·m⁻¹·K⁻¹, with a negligible electronic contribution, confirming that heat transport is dominated by lattice phonon scattering. As a result of the combined increase in electrical conductivity and intrinsically low thermal conductivity, the dimensionless figure of merit (ZT) increased continuously with temperature and reached 0.17 at 673 K. These results demonstrate that the MA–HP route provides an effective and scalable strategy for producing phase-pure Cu₃BiS₃ with controlled microstructure and reproducible thermoelectric performance. Full article

The integration of wind energy into urban environments is constrained by low wind speeds, high turbulence, and the recurrent negative torque experienced by lift-driven vertical-axis wind turbines (VAWTs). This study specifically evaluates a straight-bladed H-Darrieus rotor equipped with a single upstream passive flat-plate deflector for the wind regime measured on the campus of the Universidad Nacional Mayor de San Marcos (Lima, Peru). A three-dimensional transient CFD model using the SST k–ω turbulence model was applied to compare the baseline rotor and the deflector-assisted configuration under identical operating conditions; DMST calculations were used only as a low-order cross-check for the bare rotor performance trend, not as a substitute for experimental validation. The deflector was selected after a geometric sensitivity assessment and positioned at 30° relative to the incoming flow, with a span equal to the rotor height and a length comparable to the rotor diameter. At TSR = 2.5, the maximum power coefficient increased from 0.4459 for the bare rotor to 0.6153 with the deflector, equivalent to an improvement of approximately 38%. Velocity and pressure fields show that the deflector accelerates the flow toward the advancing blade while shielding the returning blade, thereby reducing adverse torque and smoothing cyclic torque fluctuations. The results define the applicability of the proposed passive device for low-to-moderate urban wind environments with a dominant wind sector and provide a reproducible numerical basis for subsequent wind-tunnel and field validation. Full article

Niche and interspecific association are important components of community ecology and are of great significance for revealing the mechanisms of community assembly and its stability. In this study, the woody plant communities of Betula platyphylla Sukaczev forests in the Niyang River Basin of southeastern Qinghai–Tibet Plateau were taken as the research object. The niche, interspecific association, and community stability of dominant tree species in B. platyphylla forests were analyzed using the Levins index (B_L), Shannon index (B_S), Pianka index (O_ik), Schoener index (C_ik), variance ratio (VR), chi-square test, association coefficient (AC), Spearman rank correlation, and M. Godron stability methods. The results showed that a total of 71 woody plant species were recorded across 48 plots, mainly belonging to Rosaceae, Ericaceae, and Caprifoliaceae. B. platyphylla, Quercus aquifolioides Rehder & E. H. Wilson, Sorbus rehderiana Koehne, and Berberis gyalaica Ahrendt had relatively large niche breadths, indicating strong resource utilization ability and a wide range of spatial adaptation. They were the main constructive species and dominant species of B. platyphylla forest communities in this basin. The overall niche overlap of woody plant communities was relatively low, indicating relatively obvious differentiation in resource utilization among different species. Interspecific association analysis showed that the dominant species in the tree layer exhibited an overall significantly positive association, whereas those in the shrub layer exhibited an overall non-significantly positive association. The associations between species pairs were mainly non-significant, and the overall interspecific association was weak. Most species showed a relatively independent distribution pattern, reflecting weak interspecific competition within the community. Community stability analysis showed that the Euclidean distance between the tree layer and the theoretical stability point (20, 80) was 20.17, whereas that of the shrub layer was 27.98, indicating that the tree layer was more stable than the shrub layer. Overall, the community may not yet have reached a fully stable state. The results provide important references for biodiversity conservation, vegetation restoration, and sustainable forest management in alpine canyon ecosystems. Future studies should incorporate environmental factors such as soil properties and hydrothermal conditions to further reveal the ecological mechanisms driving community succession and stability. Full article

Alpine shrub meadows on the Qinghai–Tibet Plateau are key warm-season pastures that support pastoral production and ecosystem stability in fragile high-elevation regions. Due to low temperatures, short growing seasons, and slow vegetation recovery, these pastures are highly sensitive to inappropriate grazing management. However, the effects of different grazing–rest time configurations on plant community composition and leaf functional traits in alpine shrub meadows remain insufficiently understood. In this study, we evaluated five grazing treatments in an alpine shrub meadow in Sunan County, central–eastern Qilian Mountains: 10 days grazing–20 days rest (T1), 15 days grazing–15 days rest (T2), 20 days grazing–10 days rest (T3), continuous grazing (CG), and grazing exclusion (CK). In the third year of treatment implementation, we measured the community diversity, species importance values, and leaf functional traits of four dominant species: Elymus nutans, Carex tibetikobresia, Oxytropis kansuensis, and Bistorta vivipara. T1 and T2 significantly increased species richness, Shannon–Wiener diversity, and Simpson diversity compared with CG and CK. NMDS and PERMANOVA further showed significant differences in overall community composition among grazing treatments. Grazing generally reduced the leaf length, leaf width, and leaf area, whereas T2 showed relatively stronger leaf recovery among grazing treatments. Specific leaf area, specific leaf weight, and leaf length–width ratio showed higher variability and calculated plasticity than leaf thickness and leaf dry matter content, suggesting that resource-acquisition and morphological traits were more responsive to grazing than conservative structural traits. The coefficient of variation of leaf traits was positively associated with the plasticity index, although this association should be interpreted cautiously because both indices were calculated from the same underlying trait dataset. Overall, under the conditions of this three-year, single-site experiment and a target moderate grazing intensity, the 15-day grazing–15-day rest regime performed best among the tested treatments. This regime may provide a practical reference for rotational grazing management in similar warm-season alpine shrub meadows, but its broader applicability requires further validation across different grassland types, grazing intensities, climatic conditions, and longer monitoring periods. Full article

Background: Chest radiography remains central to post-procedural assessment of central venous catheter (CVC) placement in intensive care units. Multimodal large language models (MLLMs) can process medical images, but their reliability for practical radiography tasks remains uncertain. This study assessed the diagnostic performance of MLLMs and intensivists for CVC access classification, CVC tip assessment, and pneumothorax-related radiographic findings. Methods: In this retrospective diagnostic performance study, consecutive portable anteroposterior chest radiographs obtained after CVC placement in adult critically ill patients were independently evaluated by four intensivists and five MLLMs. A radiologist consensus served as the reference standard. Interobserver agreement and diagnostic performance were assessed using Fleiss’ kappa, Gwet AC1, Cohen’s kappa, accuracy, sensitivity, specificity, precision, F1 score, balanced accuracy, and Matthews correlation coefficient. Results: The final cohort included 183 unique radiographs. Intensivist reviewers showed high performance for CVC access classification but lower and more heterogeneous performance for CVC tip-position assessment. Among MLLMs, CVC access accuracy ranged from 0.339 to 0.874, whereas CVC tip assessment was dominated by almost universal classification of tips as appropriate, with near-zero specificity and chance-level balanced accuracy. For pneumothorax-related findings, all MLLMs classified every case as negative. Intensivist reviewers had higher balanced accuracy than MLLMs for CVC access classification (difference, 0.420; 95% CI, 0.349–0.490; p < 0.001) and CVC tip assessment (difference, 0.247; 95% CI, 0.205–0.290; p < 0.001). Pneumothorax analyses were exploratory because only five positive cases were present. Conclusions: The evaluated MLLMs showed unreliable diagnostic performance compared with experienced intensivists. Apparent performance was influenced by class imbalance and dominant-response behavior, supporting cautious task-specific validation and complete diagnostic performance reporting. Full article

The quantitative identification of heavy metal sources is essential to clarify their relationships with ecological and health risks. This study focused on the Manghe Watershed in Jiyuan City, Henan Province, China, integrating the Positive Matrix Factorization (PMF) model, ecological risk index (RI), and health risk assessment (HRA) to construct a coupled PMF-RI/PMF-HRA framework to quantify source-specific risk contributions and propose targeted mitigation strategies. Key findings included: (1) Among the 121 surface soil samples, Cr and Ni showed natural background levels, while Cd, Pb, Hg, Zn, As, and Cu exceeded regional backgrounds by 1.63–33.65 times with anthropogenic-driven spatial heterogeneity. (2) The PMF identified four sources: natural–agriculture mixed (42.65%), the main contributor to Cr, Ni, As, and Cu; industrial activity (24.99%), the primary source of Cd and Zn; traffic–agriculture mixed (20.99%), primarily emitting Pb and As; and coal combustion (11.36%), dominating Hg emissions. (3) Ecological and health risks were governed by heavy metal toxicity and exposure pathways rather than mere concentration levels. Specifically, industrial sources (Cd, Zn) should be prioritized for ecological risk control, whereas natural–agricultural mixed sources (As, Pb, Cr) should be prioritized for health risk control. Oral ingestion was the dominant exposure pathway for both non-carcinogenic risk and carcinogenic risk in children, with the natural–agricultural mixed source contributing the most to this pathway. (4) The total carcinogenic risk (TCR) for children was 1.17 × 10⁻⁴, which exceeds the commonly accepted unacceptable threshold of 1 × 10⁻⁴, indicating a potential carcinogenic concern. (5) The PMF-RI and PMF-HRA frameworks quantitatively proved that the main sources of ecological risks and health risks may be completely different, and this phenomenon was jointly regulated by the toxicity response coefficient and exposure pathways. A “source–risk-pathway” quantitative attribution was achieved and provides clear support for targeted interventions, emphasizing source control for industrial emissions (Cd-Zn), traffic–agriculture inputs (Pb-As), and coal-derived Hg, alongside optimized agricultural practices. Full article

To investigate the association between polymorphisms in the GABRA5, SOX13, and AGL genes and growth traits in Dongfeng sika deer and to identify potential molecular markers for breeding, this study was conducted based on prior genome-wide association analysis. Based on the previous GWAS analysis of 266 Dongfeng sika deer, the SNP loci of GABRA5, SOX13, and AGL genes were detected in 36 male deer samples. The genetic parameters were calculated, and an association analysis with growth traits was carried out. Phenotypic analysis indicated that body weight and chest circumference had higher coefficients of variation than other growth traits, and body weight showed a strong positive correlation with body-slant length (r = 0.743, p < 0.01) and a moderate correlation with chest circumference (r = 0.709). A total of six SNP loci were identified, including three within GABRA5 (Chr13-8442730, Chr13-9033380, and Chr13-9045819), one within SOX13 (Chr14-5681678), and two within AGL (Chr20-66603370 and Chr20-66618510). The dominant genotypes at these loci include CG (CC), AA, CG, CC (CG), AA, and GG (GC). Linkage disequilibrium analysis revealed a relatively strong association between Chr13-8442730 and Chr13-903380 on chromosome 13. Combined genotype analysis showed that diplotype CCCGGC was associated with higher body weight and larger chest circumference than other genotype combinations. Gene expression analysis showed that the relative expression levels of GABRA5, SOX13, and AGL were lower in the low-growth group than in the high-growth group, and expression variation was also observed within groups. Overall, gene expression levels appeared to be positively associated with growth traits, with higher expression associated with improved growth performance. These findings suggest that GABRA5 and AGL may serve as candidate genes for further investigation and that the identified SNP loci may contribute to the development of molecular markers for the selection of growth traits in Dongfeng sika deer. The results provide a preliminary basis for molecular breeding and genetic improvement strategies in Dongfeng sika deer bucks and serve as an important reference for genetic improvement of growth traits in Cervidae. Full article

Methane emission reduction represents a critical pathway for near-term climate mitigation. Super-emitter control is widely recognized as the most cost-effective mitigation strategy; however, the prevalence of these sources varies significantly across countries and sectors, resulting in heterogeneity in abatement difficulty and policy priorities. In this study, we integrate recently emerging satellite-based point-source emission datasets to develop a cross-scale analytical framework that systematically characterizes methane emission rate distributions across countries and sectors. Analysis of the full Carbon Mapper dataset shows that sources exceeding 5000 kg h⁻¹ account for only 3.34% of total point sources, yet contribute more than 25.18% of total equivalent emissions. Gini coefficients range from 0.46 to 0.60 across countries, indicating pronounced inequality in emission distributions and mitigation costs. Integrating these distributional characteristics with economic capacity indicators further shows that countries with highly concentrated, high-intensity point sources—particularly oil- and gas-dominated nations such as Turkmenistan and Uzbekistan—offer the highest cost-effective mitigation potential and should be prioritized as global methane action breakthroughs. Among these, economically advanced countries are positioned to lead by demonstration, while nations with high mitigation potential but limited economic capacity represent optimal targets for international climate finance and technology transfer. These findings provide satellite-derived evidence to inform differentiated, country- and sector-specific mitigation pathways. Full article

Handheld X-ray fluorescence (HHXRF) scanners generate rapid, low-cost geochemical datasets from core and cuttings, yet their potential for quantitative reservoir characterisation remains largely unrealised, partly because standard multivariate methods are inappropriate for the compositional nature of geochemical data. Here we test, for the first time within a compositional data analysis framework, whether centred log-ratio-transformed HHXRF element compositions can simultaneously predict plug-scale porosity, directional permeability and facies in a siliciclastic reservoir in a continuously cored Brent Group well from the Northern North Sea. The cored interval was logged for facies, sampled for routine core analysis, and analysed by HHXRF at plug sample positions. Sixteen consistently detectable elements were transformed using centred log-ratios to respect the compositional nature of the data, and four Random Forest models were trained: regression models for porosity, horizontal permeability and vertical permeability and a seven-category facies classifier. Models were evaluated using out-of-bag predictions, residual analyses, class-wise reliability metrics and permutation-based variable importance. The models reproduce porosity and permeability with high coefficients of determination (R² > 0.95) and low errors relative to observed ranges and achieve facies classification with substantial agreement (κ = 0.705), with best performance in clean sandstone facies. Predictive skill is dominated by a consistent subset of elements (notably Ca, Ti, Si, V, Zn and Rb), linking bulk composition to mineralogy, depositional texture and diagenetic modification. These results demonstrate that compositional data from HHXRF alone can quantitatively recover key reservoir attributes and facies architecture at plug scale, establishing bulk geochemistry as a robust proxy for reservoir quality in quartz-rich, moderately buried siliciclastic reservoirs. The workflow provides a methodological template for integrating compositional geochemistry with machine learning in subsurface characterisation and, pending multi-well validation, offers a route to cost-effective prediction of porosity, permeability anisotropy and facies from cuttings or high-resolution core scanning. The workflow has direct application to geocellular model population in carbon and hydrogen storage sites, geothermal reservoirs and conventional hydrocarbon fields. Full article

Saline water irrigation has emerged as a promising approach to mitigate agricultural water shortages; however, its improper use may induce secondary soil salinization. In this study, saline-alkali soil collected from Hami, Xinjiang, was used to conduct a series of indoor one-dimensional vertical soil column experiments. The aim was to systematically investigate the effects of sodium humate and desulfurization gypsum on soil infiltration behavior and the distribution patterns of key cations and anions under different levels of irrigation water salinity. The results showed that sodium humate application markedly improved soil infiltration capacity, while the duration of infiltration decreased with increasing salinity. Under salinity levels of 12 and 16 g/L, the 4 g/kg sodium humate treatment exhibited the most rapid advancement of the wetting front. In contrast, desulfurization gypsum reduced infiltration rates, with the lowest infiltration observed under the 12.5 g/kg treatment at 16 g/L salinity. Under different treatments, the adjusted coefficients of determination (adjusted R²) for the Philip, Kostiakov, and Horton models ranged from 0.8450 to 0.9841, 0.9901 to 0.9989, and 0.9748 to 0.9942, respectively, while the global performance indicator (GPI) ranged from 1.619 × 10⁻³ to 5.103 × 10⁻¹, 4.998 × 10⁻⁹ to 2.166 × 10⁻⁵, and 1.505 × 10⁻⁶ to 2.438 × 10⁻⁴, respectively. These results indicate that the Kostiakov model outperformed the other models in terms of fitting accuracy and overall performance for describing the soil infiltration process. In addition, sodium humate generally increased the sorptivity parameter S in the Philip model and the empirical coefficient K in the Kostiakov model, whereas desulfurization gypsum showed the opposite trend. In terms of salt regulation, sodium humate demonstrated optimal desalination performance at application rates of 6–8 g/kg under low salinity and 4–6 g/kg under high salinity conditions. Conversely, excessive gypsum application tended to exacerbate salt accumulation, although a moderate dosage (5 g/kg) effectively limited the downward migration and accumulation of Na⁺ and Cl⁻. These two ions were identified as the dominant contributors to soil salinization, showing strong positive correlations with soil salt content (SSC), sodium adsorption ratio (SAR), and exchangeable sodium percentage (ESP). In contrast, Ca²⁺, Mg²⁺, and HCO₃⁻ played beneficial roles in alleviating sodicity through ion exchange and buffering mechanisms. Overall, sodium humate enhanced infiltration and facilitated salt leaching in the upper soil layers under saline irrigation conditions. Although desulfurization gypsum reduced infiltration and increased overall salt content, it contributed to mitigating Na⁺ accumulation in deeper soil profiles. These findings highlight the critical importance of selecting appropriate soil amendments and optimizing their application rates to improve saline water use efficiency and promote sustainable management of saline-alkali soils. Full article

The highly nonlinear aerodynamic interference in tandem-airfoil configurations significantly hinders the precise exploitation of their aerodynamic potential. To address this issue, this study establishes a high-fidelity computational fluid dynamics benchmark. A high-quality sample set is constructed using Latin hypercube sampling combined with an intra-layer replacement strategy. Subsequently, a Gaussian process surrogate model and Bayesian optimization are employed to maximize the total system lift coefficient across a four-dimensional design space comprising longitudinal and vertical separations, fore airfoil angle of attack, and angle of attack difference. Global sensitivity analysis indicates that longitudinal separation dominates the interference modes. Optimization reveals a distinct mode switching phenomenon using a longitudinal separation of twice the chord length as the critical threshold. In the close-coupled configuration, a negative optimal angle of attack difference enhances the slot effect and upwash induction, thereby delaying rear airfoil stall and achieving synergistic lift enhancement. Conversely, in the distant-coupled configuration, the system transitions to a decoupled compensation mode, where a positive angle of attack difference compensates for the effective angle of attack loss induced by wake downwash. This research elucidates the competitive mechanisms between inter-airfoil slot flow and wake interference, providing a theoretical reference for the aerodynamic layout optimization of tandem-airfoil aircraft. Full article

Given the challenges posed by high groundwater levels, thick sand layers, and strong permeability in water-rich sandy strata, cut-off walls often fail to fully isolate the hydraulic connection between the inside and outside of a foundation pit. As a result, dewatering inside the pit—especially from confined aquifers—can cause significant external groundwater drawdown and subsequent ground settlement. Using a deep excavation conducted in Xiamen as a case study, this study developed a two-dimensional hydro-mechanical coupled finite element model to systematically investigate the effects of various dewatering scenarios and soil permeability coefficients on surface settlement around the pit, and to reveal settlement patterns induced by dewatering and excavation in such strata. Field monitoring data were incorporated to validate the numerical model, ensuring accuracy and reliability. Key findings include the following: (1) Dewatering contributes to over 76% of the total settlement at each stage, with confined drawdown being the dominant factor, implying that dewatering optimization should take priority over controlling excavation rate. (2) Under confined dewatering, the settlement influence zone extends beyond 80 m, far exceeding the extension caused by excavation alone; thus, monitoring and protection ranges must be adjusted dynamically. (3) The horizontal permeability of sand shows a nonlinear positive correlation with settlement, and this sensitivity grows with depth, highlighting the need for accurate permeability determination and stricter controls in deep excavations within water-rich sand layers. From an engineering perspective, these findings underscore the importance of prioritizing confined aquifer dewatering management, dynamically expanding settlement monitoring zones, and rigorously characterizing permeability profiles to mitigate excessive ground settlement and protect adjacent infrastructure. Full article

Understanding the pore structure of high-rank coals is essential in evaluating gas storage and transport. Here, twelve semianthracite samples from the early Permian Shanxi Formation were investigated by proximate analysis, optical microscopy, low-temperature N₂ adsorption, and fractal analysis, coupled with diffusion coefficient modeling. The coals exhibit diverse pore types (plant-cellular, interparticle, and dissolution pores) shaped by coalification and minerals and show Type IV (a) isotherms with H4 hysteresis loops, indicating complex pore networks. Pore-size partitioning reveals that mesopores and macropores dominate total pore volume, whereas mesopores contribute most of the specific surface area. The pore structure exhibits strong fractal characteristics with an average comprehensive fractal dimension (F_c) of 2.628. The calculated gas diffusion coefficient decreases monotonically with increasing pressure from 1 MPa to 5.8 MPa, with a more pronounced decline at low pressure, indicating a clear pressure-dependent attenuation effect. Diffusion capacity is weakly related to average pore diameter but shows positive correlations with total pore volume and, particularly, macropore volume. Multiple linear regression further demonstrates that pore volume structure is the dominant control on diffusion under both low- and high-pressure conditions, with the relative importance ranked as macropores > mesopores > micropores. Macropores provide the main low-resistance transport framework, mesopores serve as transitional pathways linking storage and transport domains, whereas micropores mainly contribute to gas storage and may even suppress apparent diffusion when overly developed. These results reveal a clear functional differentiation of multiscale pore systems and highlight that gas migration in semianthracite is jointly governed by pore size distribution, connectivity, tortuosity, and fractal network topology. Full article

The bond strength of corroded reinforced concrete (CRC) structures is critical for structural safety and long-term durability. However, the corrosion-induced bond degradation process is influenced by multiple, coupled factors and exhibits complex, nonlinear behavior, making it difficult for traditional theoretical models to provide accurate predictions. To address this challenge, this study proposes a novel, unified prediction framework based on machine learning techniques. A total of 391 experimental datasets were collected and compiled, covering key parameters including bond strength, reinforcing bar diameter, yield strength, concrete cover thickness, concrete compressive strength, mass loss rate due to corrosion, and the presence of stirrups. Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) algorithms were employed to develop predictive models for bond strength. Model training and testing were performed using 10-fold cross-validation. Furthermore, the SHapley Additive exPlanations (SHAP) approach was introduced to enhance model interpretability and quantitatively assess the influence of each input feature, revealing that mass loss rate and bar diameter are the dominant factors. This study effectively bridges the research gap between high-precision black-box algorithms and the need for physical interpretability in engineering. The results demonstrate that (1) the proposed XGBoost model significantly outperforms traditional empirical formulations, achieving a high coefficient of determination (R² = 0.893) and a much lower coefficient of variation (25.85%) on the testing set, and (2) the SHAP analysis reveals that the machine learning predictions are highly consistent with established physical mechanisms, successfully capturing the negative impact of splitting tensile stresses caused by rust expansion and the positive confinement effect of stirrups. Overall, the proposed models demonstrate superior accuracy, robustness, and generalization capability, providing an effective tool and theoretical basis for evaluating bond behavior and designing durable CRC structures with broad engineering applicability. Full article