网络首发
Research on the migration mechanism and remediation of groundwater contaminants in leakage area of Zhuluo product oil pipeline
Zhang Jingfei;Duan Shengcong;Zhu Changcheng;Zhang Yu;Zhong Wei;Wang Chenyu;A coupled numerical model was developed to simulate the full-process migration of petroleum contaminants in soil and groundwater systems following a refined oil pipeline leakage. This model was established using the COMSOL Multiphysics platform and integrated variably saturated flow in the vadose zone with saturated solute transport in the aquifer. The governing equations included the Richards equation for unsaturated flow and the advection–dispersion equation for solute transport. The model accounted for gravity-driven infiltration， capillary retention， mechanical dispersion， and molecular diffusion. A layered heterogeneous hydrogeological domain was constructed， with hydraulic conductivity and porosity varying spatially. A time-dependent concentration source represented continuous leakage， while atmospheric infiltration was assigned at the ground surface. No-flow boundaries were applied laterally， and a constant-head boundary was defined at the downstream aquifer. The model was solved using a finite element scheme with adaptive time stepping to ensure numerical stability and convergence. Simulation results indicate that contaminant migration in the vadose zone is controlled by both gravity and capillary forces， significantly delaying contaminant breakthrough， which occurs after approximately 3.2 years， with a breakthrough concentration of 0.258 mol/m³. The retention time in the vadose zone is about 1800 days， during which it acts as both a transport pathway and a buffering medium， regulating contaminant release into the aquifer. Once contaminants enter the aquifer， advection dominates transport， while hydrodynamic dispersion and heterogeneity influence plume deformation. The plume elongates along the flow direction and spreads laterally， reaching an area of 0.76 km2 at 1800 days， which aligns well with field measurements. Additionally， a pump-and-treat optimization framework was developed based on plume delineation， testing various well configurations， pumping rates， and spatial layouts. Results demonstrate that well placement significantly impacts capture efficiency， with wells aligned along the plume centerline improving hydraulic control. The optimal configuration includes seven extraction wells， each operating at 300 m³/d， achieving effective plume containment and contaminant removal. This model serves as a quantitative tool for groundwater monitoring design and remediation optimization in petroleum-contaminated sites.
Effects of caved coal and rock on adsorption and release of characteristic pollutants in mine water from the Pingdingshan Mining Area
Lu Kaixuan;Luo Qing;Sun Daoning;Xu Lianman;This study examines the influence of caved coal-rock masses on the migration and release of characteristic pollutants， such as Fe and Pb， in mine water. First， the composition， structure， and surface element distribution of coal and gangue were comprehensively characterized using SEM mapping， XRD， XRF， and IR techniques. Subsequently， static immersion experiments were conducted to assess various influencing factors， including different pH values， immersion times， and particle sizes of coal and rock. The migration and release patterns of Fe and Pb ions were determined using ICP analysis. Finally， low-temperature nitrogen adsorption experiments were performed to elucidate the mechanisms underlying the migration and release of these pollutants. The results indicate that: (1) In deionized water， gangue is the primary source of Fe ions， with the amount released influenced by particle size and immersion time; specifically， smaller particle sizes lead to higher Fe release during the initial immersion stage. The source of Pb ions is affected by the combined action of coal and gangue; while the presence of gangue enhances Pb release， no significant relationship was found between the amount of Pb released and gangue content. (2) Under mine water conditions， coal has minimal impact on the concentration of Fe ions in the water， and other ions present inhibit the release rate of Fe from gangue; Pb release is jointly regulated by both coal and gangue. (3) Low-temperature nitrogen adsorption experiments reveal that the migration and release of metal ions such as Fe and Pb in mine water are significantly influenced by the pore structures of coal and gangue. In summary， controlling iron pollution in mine water should focus on the particle size of gangue and the inhibitory effects of other ions on iron release， whereas lead pollution control requires a comprehensive consideration of the interactions between coal and gangue. This study provides a theoretical basis and technical support for the in-situ purification and resource utilization of mine water.
Four-party evolutionary game study on CDW recycling governance
Zhang Min;Cao Yue;To address the governance dilemma of Construction and Demolition Waste (CDW) recycling—characterized by high output， low recycling rates， and insufficient synergy—this paper constructs a four-party evolutionary game model within the framework of “government， recycling enterprises， the public， and academic and research institutions.” The model captures the coupling mechanisms of policy incentives， market feedback， public participation， and technological progress. Under the assumption of bounded rationality， the replicator dynamic equations are derived. Through numerical simulation and sensitivity analysis， we compare the effects of different regulatory intensities and incentive combinations on the recycling rate， convergence speed， and fiscal efficiency. Additionally， a multi-scenario heatmap simulation is conducted using two-parameter scanning with a fixed third parameter to examine the synergistic effects of key parameters. The findings indicate that a single policy of high subsidies tends to crowd out regulations and incentivize free-riding， making it difficult to achieve long-term performance. Reducing the costs associated with public participation generates a significant amplification effect through the transmission chain of “source separation quality → reduction in enterprise sorting costs → increase in enterprise net benefits”. Targeted research incentives for academic and research institutions enhance the technology success rate and expand the interval of high-performance equilibrium. The effectiveness of subsidies and fines is fundamentally constrained by regulatory costs; only when these three elements are proportionately balanced can the system transition from a low-level lock-in to a high-recycling equilibrium. Moreover， there exists a clear “minimum coupling threshold” in the technology innovation chain: a positive feedback loop can only be achieved when research incentives， R & D success probability， and enterprise absorptive capacity work in concert. Accordingly， this paper proposes a policy combination path that starts with rigid regulation and integrates performance-linked incentives， research support， and public participation， thereby providing theoretical foundations and decision-making references for the governance of construction and demolition waste recycling.
Study on preparation and properties of fluorine-free surfactant compound foam extinguishing agent
Qian Xinming;Hou Yunhe;Zhang Jianqi;Hao Wenhao;Li Minghui;Li Fangzhou;Wu Xingtong;Yu Lin;Wang Kuo;To improve the foam performance of clean fluorine-free fire extinguishing agents while ensuring environmental friendliness， this study adopts binary and multicomponent compound methods combined with orthogonal analysis and grey relational analysis to test parameters of fluorine-free surfactant systems such as foam volume， liquid drainage time， surface tension， viscosity， extinguishing time， and foam coarsening. Numerical calculations were obtained through quantitative analysis to optimise the compound formula， and its property parameters were compared with commonly available superior fire extinguishing agents on the market. The results indicate that α-olefin sulfonic acid salts（AOS）， sodium dodecyl sulfate（SDS）， and alkyl polyglycoside（APG）as foaming agents possess superior performance and synergistic effects; when blended in optimal ratios， the foam volume can reach 375 mL and surface tension is 29.1 mN/m. Glycerol， urea， and Modified polyethoxylated silicone（FM-550）as additives can effectively improve the comprehensive performance of foam extinguishing agents. Through comprehensive statistical analysis of the foam volume， 50% liquid drainage time， surface tension， and viscosity of 16 compound groups using orthogonal analysis and grey relational analysis， the optimal mass fractions are calculated to be 0.2%， 0.3%， and 0.2% respectively. Fire extinguishing performance tests verified that the obtained formula can quickly and efficiently extinguish 34 B and 55 B oil pan fires， with extinguishing times of 28.7 s and 36.7 s respectively. Compared with domestically and internationally leading fluoroprotein foam extinguishing agents and AB-type foam extinguishing agents， the formula obtained in this study often has superior comprehensive performance indicators， with a foam volume of 380 mL， 50% liquid drainage time of 263 s， surface tension of 23.4 mN/m， and viscosity of 2.77 mP·s. By comparing the photographed foam coarsening images， the foam produced by this extinguishing agent is characterised by very uniform and fine bubbles， with a low level of coarsening over 20 minutes.
Research on robust photovoltaic system fault diagnosis model using RCET–TGatedGCN
Shen Chunming;Zhao Wen;Zhao An;Wang Zhicheng;Han Fangwei;Photovoltaic (PV) fault diagnosis is essential for the safe and reliable operation of PV systems. However， in practical deployments， diagnostic performance is often compromised by measurement noise and data loss introduced throughout the sensing， transmission， and acquisition chain. To address this challenge， this paper proposes a physically constrained multi-relational graph convolutional network， named RCET–TGatedGCN (Reliability-Constrained Edge Typing–Gated Graph Convolutional Network with Temporal Adaptation)， designed for robust PV fault diagnosis under disturbed data conditions. First， a Reliability-Constrained Edge Typing (RCET) strategy is formulated to construct a layered graph topology that distinctly separates highly reliable physics-driven relations from disturbance-sensitive statistical correlations. This design minimizes the propagation of spurious dependencies and enhances the physical interpretability of graph modeling. Second， we develop a feature-quality-aware multi-scale anti-interference mechanism comprising Type-Scaling and Type-Gate modules. The Type-Scaling module suppresses noise-sensitive feature components prior to message passing， thereby improving the quality of relation-specific node representations. The Type-Gate module adaptively adjusts the aggregation weights of different relation channels based on feature reliability， allowing the model to prioritize more trustworthy information amidst stronger interference. Moreover， sparsity regularization and DropEdge techniques are introduced for the noisy auxiliary relations to mitigate overfitting to unstable correlations and enhance structural robustness. Experiments are conducted on a MATLAB/Simulink-based dataset featuring a 3×3 PV array with 15 operating states， in which graded Gaussian noise and random missing values simulate realistic disturbances. Under standard interference conditions， the proposed method achieves a macro-average F1 score of 91.73%， representing an improvement of 3.58 percentage points over the baseline GCN. In full-spectrum stress tests， the macro-average F1 score remains at 71.69% even under extreme interference， with a relative performance decay controlled at 21.85%. Additionally， in LOCO cross-condition evaluations， the generalization gap is limited to 2.61%， demonstrating strong robustness to covariate shifts and superior cross-condition generalization capabilities. These results confirm that the proposed method effectively enhances both physical consistency and diagnostic reliability for PV systems operating in complex and disturbance-rich environments.
Fire extinguishing efficiency of three fire extinguishing systems in full-scale cable tunnel fires
Yan Long;Wang Junyi;Xu Zhisheng;Li Xin;To evaluate the effectiveness of fire extinguishing systems in cable tunnel fires， this paper combines cone calorimeter testing with full-scale physical experiments conducted in cable tunnels. It systematically reveals the combustion characteristics and evolution patterns of cable fires， analyzes the fire suppression efficacy during the transition from the initial growth phase to the rapid development phase for three extinguishing systems—liquid nitrogen， high-pressure fine water mist， and ultra-fine dry powder—and elucidates the extinguishing mechanisms of these systems based on the results obtained. Both the cone calorimeter and full-scale cable tunnel tests confirm that cable materials exhibit a strong tendency for secondary combustion and possess extremely high self-sustaining combustion capabilities. In the absence of external thermal radiation， the Peak Heat Release Rate (PHRR1 and PHRR2) of the cable reaches as high as 145.1 kW/m2 and 135.1 kW/m， respectively. The results from the full-scale fire extinguishing tests indicate that when the liquid nitrogen extinguishing system is positioned directly above the cable bridge， it demonstrates the best fire control efficacy， extinguishing the fire in only 184 s with a cooling rate of up to 3.50 ℃/s. In contrast， the high-pressure water mist system takes approximately 250 s to extinguish the fire， exhibiting progressive suppression characteristics. Although the ultra-fine dry powder system can suppress open flames， it is prone to rekindling after extinguishment due to its low cooling efficiency. The differences in fire extinguishing efficiency among the three systems for cable tunnel fires primarily stem from their distinct dominant mechanisms. Liquid nitrogen penetrates deep into the fire source effectively due to rapid phase-change heat absorption and suffocation resulting from its extremely low temperature. The high-pressure water mist relies on droplet vaporization for heat absorption and local suffocation， but its efficacy is limited by the droplet penetration force and vaporization rate. The ultra-fine dry powder mainly functions through chemical inhibition to disrupt the chain reaction， making it challenging to effectively dissipate deep heat; furthermore， its particles fail to penetrate the cable bundle adequately， resulting in a high risk of reignition.
Study on neurophysiological mechanisms of cognitive fatigue in night shift miners based on EEG-fNIRS
Chen Ying;Lu Jikun;Lu Caiwu;Gu Qinghua;Jiang Song;Zou Yuan;To objectively evaluate the neurophysiological mechanisms underlying cognitive fatigue among underground miners during night-shift operations， this study proposes a multimodal assessment framework combining Electroencephalography and Functional Near-Infrared Spectroscopy. To ensure ecological validity， an underground mine roadway simulation cabin was constructed. A rigorous self-controlled experimental design was implemented to compare the neurocognitive states of 30 professional miners before and after their night shifts. Utilizing portable multimodal EEG-fNIRS equipment， researchers synchronously acquired continuous brain electrophysiological signals and cerebral hemodynamics data while participants executed the Psychomotor Vigilance Task. Subjective fatigue levels were concurrently evaluated using standardized questionnaires to establish a comprehensive psychological baseline. The comprehensive data analysis yielded several critical findings. Behaviorally， the Mean Reaction Time during the PVT was significantly prolonged after the night shift， demonstrating a remarkable increase of 97.5%. This objective decrement was consistent with significantly elevated scores on the subjective Karolinska Sleepiness Scale. Neurophysiologically， EEG data revealed that the spectral power of theta (θ) and alpha (α) frequency bands in frontal brain regions significantly increased under fatigue conditions， with the theta/alpha power ratio index exhibiting a marked elevation. Furthermore， fNIRS data demonstrated that the task-evoked activation response of oxygenated hemoglobin in the prefrontal cortex was substantially attenuated during the fatigued state. Correlation analyses highlighted that theta wave power variations were significantly and positively correlated with MRT prolongation， whereas changes in oxygenated hemoglobin displayed a significant negative correlation with MRT. These findings comprehensively demonstrate that intensive night-shift work induces a profound "neurovascular decoupling" within the prefrontal cortex of miners. This decoupling is characterized by simultaneous neural firing suppression and a fundamental mismatch between blood oxygen metabolic supply and demand， serving as the primary physiological mechanism driving cognitive decline. Ultimately， this multimodal framework effectively elucidates night-shift fatigue mechanisms， providing a robust objective basis for fatigue assessment and intelligent safety early-warning systems for underground mining personnel.
Research on identifying and correlating risk factors of construction supervision entities based on text mining
Dai Chengliang;Yang Wentao;Ren Yongbo;Gao Jindong;This study aims to identify the risks in the work of regulatory bodies in building construction and reveal the internal correlations among various risk factors. Firstly， a text corpus composed of 247 work safety accident investigation reports of building construction is constructed. With the adoption of text mining technology， word segmentation is implemented on the text corpus based on professional dictionaries， and the segmentation results are optimized by integrating the thesauri of new words， stop words and synonyms. The TF-IDF method is applied to extract 38 keywords， which are divided into two dimensions of "subsystem" and "deviation". According to various descriptions of the risks of regulatory bodies in building construction reflected in accident texts， 22 risk factors of regulatory bodies are identified via expert seminars and further categorized into three dimensions: inadequate regulatory construction， vacant inspection and guidance， and ineffective work implementation. Construct a co-occurrence matrix based on the Boolean dataset of risk factors， calculate the degree centrality of each factor， and identify key risk factors. The Apriori algorithm is employed with the minimum support set to 0.1， the minimum confidence set to 0.5， and the minimum lift set to 1.1 to mine association rules of risk factors and clarify the correlation relationships between each factor. The results demonstrate that C4 (failure to implement hidden danger investigation and treatment)， C3 (inadequate publicity， education and training)， A2 (failure to carry out work in accordance with laws and regulations)， and B3 (perfunctory law enforcement and inspection) have closer connections with other factors in the risk factor network. These are the pivotal risk factors prone to problems in the operation of building construction regulatory bodies， and targeted risk prevention and control shall be formulated correspondingly.Association rules with high support and high confidence can effectively reflect the typical interaction relationships among risk factors in the daily work of regulatory bodies. Risk factors such as A9 (imperfect supervision system)， B4 (inadequate supervision and guidance)， B5 (inadequate patrol and inspection)， and C6 (failure to adopt effective rectification measures) often produce coupling effects with other risk factors in accidents. The occurrence of B1 (negligence and dereliction in safety supervision)， A6 (formalized supervision work)， B6 (absence of supervision and inspection) and other factors will be accompanied by the emergence of more adverse factors. When other risk factors occur， A2， A9 (imperfect supervision system)， B3 (perfunctory law enforcement and inspection)， B4 and other factors are more likely to arise as combined associated factors.
Fire evolution characteristics and residual load-bearing capacity of Qing-style Five-Step Dou-gong brackets treated with transparent fire-retardant coatings
Yan Long;Zhao Zhijing;Gong Chenfei;Li Xin;Yi Liang;To enhance the fire resistance of traditional timber structures， this study systematically investigates the fire evolution behavior， charring characteristics， and residual load-bearing capacity of full-scale Qing-style Five-Step Dou-gong brackets protected with a transparent fire-retardant coating. Experiments were conducted using an ISO 9705 standard fire test platform in conjunction with a TAW–2000 multi-functional mechanical testing system. The results demonstrate that the application of the transparent fire-retardant coating significantly improves the fire performance of the Dou-gong components. Specifically， the coating delayed the initial ignition time by 60 s and reduced the peak flame height and peak heat release rate by 15.8% and 23.0%， respectively， compared to unprotected brackets. During combustion， the coating facilitated the formation of a dense， expanded char layer on the wood surface， effectively inhibiting heat and mass transfer， thereby controlling the charring process and minimizing mass loss. For instance， the most severely damaged component， the Zhengxin Guagong， exhibited reductions in charring depth and mass loss rate of 31.9% and 11.1%， respectively. Components oriented along the Qiaotou direction， which experienced weaker heat feedback， demonstrated even more pronounced improvements， with charring depth reduction rates exceeding 54.0%. Post-fire mechanical tests revealed that the protected Dou-gong brackets retained superior load-bearing capacity， with the maximum residual load increasing from 32.4 kN to 49.6 kN and the corresponding axial displacement rising from 47.6 mm to 82.3 mm. The load-displacement curve of the protected specimens was smoother and more stable， indicating improved structural integrity and resistance to fire-induced damage. In conclusion， the transparent fire-retardant coating not only effectively retards the combustion process and mitigates thermal degradation but also significantly enhances the post-fire mechanical performance of Dou-gong brackets. This provides a reliable technical approach for the fire protection of valuable historic timber structures. The findings offer crucial experimental data and theoretical support for performance-based fire safety design and the conservation of ancient wooden buildings.
Research on multiple driving paths of successful geological hazards avoidance: Based on the three-dimensional framework of “agent–technology–institution”
GUO Haixiang;ZHENG Runjie;LIU Yaning;HE Xinyu;DAI Shanshan;This study aimed to identify the key antecedent conditions of successful geological hazard avoidance and to explain how different conditions combined into distinct pathways. The analysis focused on successful avoidance as the outcome variable. It examined how agent-related， technology-related， and institution-related factors interacted in actual hazard response processes. It also sought to move beyond single-factor explanations and to reveal the configurational mechanisms behind successful avoidance.The study selected 50 typical cases of successful geological hazard avoidance in China from 2020 to 2024. It used a hybrid design that combined grounded theory and crisp-set qualitative comparative analysis (csQCA). First， the study coded the case materials through open coding， axial coding， and selective coding. It extracted 78 initial concepts， grouped them into 16 initial categories， and further summarized them into 6 main categories. On this basis， it developed an Agent–Technology–Institution (ATI) analytical framework. The six main categories were then transformed into csQCA condition variables， including grassroots hazard identification， public emergency literacy， meteorological risk warning， monitoring-device warning， source risk control， and integrated governance. The study then conducted a necessity test and a configurational analysis to identify the pathways leading to successful avoidance.The results show that no single condition independently determines successful geological hazard avoidance. Successful avoidance is the product of multiple conditions acting together. The analysis identifies seven effective configurations. These configurations can be grouped into four driving patterns: agent–institution synergy， technology–institution synergy， agent–technology–institution synergy， and institution-led. Different patterns correspond to different hazard contexts. In some cases， grassroots identification and public response trigger avoidance. In other cases， technical warning systems provide the main trigger. In more complex cases， agents， technologies， and institutions work together across risk detection， warning verification， and evacuation execution. Institutional arrangements play a particularly important coordinating and supporting role across the different pathways.The main originality of this study lies in three aspects. First， it treats successful avoidance， rather than disaster occurrence， as the core outcome. Second， it integrates grounded theory and csQCA to connect factor extraction with pathway identification. Third， it builds and tests an ATI framework that explains successful geological hazard avoidance from a configurational perspective rather than from a net-effects perspective. These findings provide a clearer explanation of how different conditions jointly produce successful avoidance in practice.