China has a large stockpile of iron ore tailings. Long-term accumulation not only occupies substantial land resources, but also poses environmental risks such as dust dispersion and heavy metal migration. Meanwhile, the production of traditional cementitious materials, such as cement, is energy-intensive and emits large amounts of carbon dioxide. This hinders the development of green and low-carbon construction. Utilizing iron ore tailings as cementitious materials offers an effective solution to ease both resource and environmental pressures. This paper systematically reviews the particle characteristics, chemical and mineral compositions, and potential cementitious properties of iron ore tailings. It summarizes the mechanisms of activity enhancement through mechanical activation, chemical stimulation, and thermal treatment. The mechanical properties, durability, and hydration products of iron ore tailings-based cementitious materials under different activation methods are discussed. The synergistic reaction potential of iron ore tailings with slag and fly ash in alkali-activated systems is highlighted. The advantages and challenges of applying iron ore tailings in the development of green construction materials are analyzed. Finally, the paper proposes that future research should focus on mineral reaction mechanisms, combined activation strategies, and multi-scale performance evaluation systems, to promote the efficient utilization and engineering application of iron ore tailings in low-carbon cementitious materials.

To promote large-scale and high-value resource utilization of iron tailings sand in the multi-layer structure of asphalt pavement for highways in Xinjiang, iron tailings sand asphalt mortar and mixtures were prepared. The optimum surface modification scheme of iron tailings sand was determined. The road performance of iron tailings sand asphalt mixtures under different content ratios was analyzed. The optimal content scheme of iron tailings sand was finally recommended. The heat transfer characteristics of iron tailings sand asphalt mixtures in heating and cooling environments were clarified. The influence laws of different factors on the damage healing performance of iron tailings sand asphalt mixtures were revealed. The service effect of iron tailings sand asphalt pavement was evaluated. The results show that 1% silane coupling agent can significantly improve the adhesion between iron tailings sand and asphalt, and the optimal modification content of modified iron tailings sand is recommended to be 60% for both AC-25 and AC-16 asphalt mixtures. Under heating conditions, the heat transfer efficiency of the iron tailings sand asphalt mixture with 60% content is higher than that of ordinary asphalt mixture, and the average temperatures of its upper and lower surfaces are 1.55 ℃ and 0.78 ℃ higher than those of the latter respectively. The light self-healing performance of iron tailings sand asphalt mixture is strongly correlated with the temperature in the crack propagation zone, and the higher the content of iron tailings sand, the better the light self-healing effect. Under the condition of 4 h light irradiation and 8 h healing, the peak load healing rate (H_P) of 60% iron tailings sand asphalt mixture reaches 54.5%. When the microwave power is 900 W and the heating time is 60 s, the H_P of the iron tailings sand asphalt mixture reaches 68.3%. In practical engineering, the average heating rate of the lower surface layer of the iron tailings sand asphalt mixture ranges from 0.157 ℃·h^-1 to 3.477 ℃·h^-1, showing better heat absorption capacity. Using it in alpine regions with large temperature differences is beneficial to improving the self-healing performance of the asphalt surface layer.

With the development of transportation infrastructure, asphalt is widely used as a key material due to its excellent performance. However, large-scale construction has led to a surge in demand for petroleum asphalt, and its non-renewability and cost pressures from price fluctuations pose challenges to industry development, making the development of environmentally friendly and cost-effective alternative materials an urgent need. Polyurethane, as a high-performance polymer material, has become an important source of solid waste due to its extensive applications in industrial and domestic fields, and its recycling has attracted significant attention. Among various methods, the alcoholysis is widely used because it can convert polyurethane into polyols for recycling through a simple process. However, the By-products generated during polyurethane alcoholysis (BPF) increase the burden of industrializing polyurethane recycling due to difficulties in their treatment. BPF is highly similar to asphalt in physical properties and apparent morphology. This paper proposes an in-depth study on using BPF as a partial substitute for asphalt. The main chemical components of BPF as ethylene oxide/propylene oxide copolyether (EO/PO copolyether) and aromatic amine compounds. Through experimental characterization and molecular dynamics methods, it was found that the blending between BPF and asphalt is primarily based on physical compatibility. The polar hydroxyl groups of the EO/PO copolyether and the amino groups of the aromatic amines in BPF form associations with asphalt molecules through non-covalent hydrogen bonding. Additionally, π-π conjugation and stacking effects occur between aromatic amines and the aromatic and colloidal components of asphalt, promoting the compatibility of BPF with asphalt. As the BPF dosage increases, the aging resistance of BPF-asphalt gradually improves, while the low-temperature crack resistance and fatigue performance show a trend of initially decreasing, then increasing, and subsequently decreasing again. In contrast, the high-temperature performance is more strongly correlated with the composition of BPF. EO/PO copolyether is the main component causing a decline in the high-temperature performance of BPF- asphalt. However, when BPF contains small-molecule polystyrene, its thermal polymerization effect becomes key to enhancing high-temperature performance. At a BPF dosage of 10%, it can serve as an effective alternative to asphalt, enabling the modified asphalt to exhibit comprehensive performance that is comparable to or even surpasses that of base asphalt in all aspects except high-temperature performance. This study validates the feasibility of using BPF as an asphalt extender, significantly promoting the resource utilization of solid waste while reducing dependence on petroleum asphalt, thereby offering both engineering applicability and sustainable development value.

To investigate the evolution laws and micro-mechanisms of the steel slag-asphalt interface under sulfate erosion conditions in saline regions of Northwest China, promoting the resource utilization of steel slag solid waste, two types of asphalt (90# base asphalt and SBS modified asphalt) and four types of aggregates (steel slag, limestone, granite, and basalt) were studied. The strength and stripping ratio of the asphalt-aggregate interface were measured using macro pull-out tests to analyze the influence of different erosion conditions on the interface properties. Nano-scale pull-out tests were conducted using molecular dynamics (MD) simulations to calculate interfacial adhesion work, nano-stripping ratio, and molecular spatial distribution, which were used to analyze the degradation mechanism of the asphalt-aggregate interface at the microscopic level. Based on Fick's second law, a damage evolution master curve model was established, characterizing the long-term performance degradation law of the asphalt-aggregate interface in sulfate erosion environments. The results show that sulfate erosion reduces the interfacial strength and increases the stripping ratio across all asphalt-aggregate combinations. As solution concentration increases and erosion time extends, the dominant failure mode shifts from cohesive failure to adhesive failure. Molecules of salt and water in the solution invade the asphalt-aggregate interface, causing the mean-square distance between these two components to increase and reducing the interfacial adhesion energy. Under various erosion conditions, the damage evolution master curves showed good agreement with the experimental data (with a goodness-of-fit larger than 0.9 in all cases), validating the effectiveness of the time-concentration equivalence principle and the master curve model. Compared to other combinations, the interfacial system formed by steel slag and SBS-modified asphalt exhibits both excellent bonding properties and strong resistance to sulfate erosion, making it an ideal choice for asphalt mixture raw materials under saline conditions. The research findings provide crucial theoretical foundations and technical pathways for enhancing the durability of asphalt pavements in saline-affected regions and promoting high-value utilization of steel slag.

To explore a high-value utilization pathway for red mud in road base courses, this paper proposes a technical solution for preparing red mud-based geopolymer-lightweight aggregate pavement base material (RGAC) through the synergistic combination of red mud-based geopolymer (RMG) and red mud-based lightweight aggregate (RA). RMG was prepared using red mud, fly ash, and ground granulated blast-furnace slag as precursors, activated by a composite alkali activator. RA was prepared using red mud and slag as the matrix through extrusion granulation, and was used to replace 4.75-9.5 mm natural aggregates at volume substitution ratios of 15% and 23%. The effects of RMG content (5%-9%) and RA content on the mechanical strength at different ages, microstructure, leaching characteristics of hazardous substances, and carbon footprint of RGAC were investigated. The results show that the 7-day unconfined compressive strength of RGAC ranges from 4.42 to 7.65 MPa, and the 28-day strength reaches 82.7%-90.7% of the 90-day strength. With increasing RMG content, the unconfined compressive strength, splitting tensile strength, and elastic modulus at all ages increase significantly, but the increase tends to flatten when RMG content exceeds 7%. When RA content increases from 15% to 23%, the unconfined compressive strength, splitting tensile strength, and elastic modulus of RGAC at different curing ages generally showed a decreasing trend. Microscopic analysis indicates that hydration products of RMG, such as N—A—S—H and C—A—S—H gels, fill pores, coat particles, and enhance interfacial bonding. When the red mud content is 40%-60% and the Si/Al ratio of the system is 1.6-2.3, the gel network is denser and the macroscopic mechanical properties are better. Excessively high red mud content (e.g., 80%) leads to insufficient active components, resulting in reduced gel formation, a loose structure, and strength deterioration. The leaching concentrations of heavy metals in all RGAC mixtures are below the Class III limits of the Groundwater Quality Standard (GB/T 14848—2017). Increasing the RMG content significantly enhances the heavy metal immobilization effect, which involves multiple synergistic mechanisms, including physical adsorption, chemical bonding, ion exchange, and precipitation reactions. Compared with cement-stabilized macadam base courses, the CO₂-equivalent emissions associated with the cementitious materials of RGAC are reduced by approximately 46%.

To elucidate the damage mechanism of steel slag pavement base materials stabilized with an all-solid-waste binder under freeze-thaw cycling, ultrasonic testing was used to determine the variation in longitudinal wave velocity before and after freeze-thaw cycling, and acoustic emission (AE) monitoring was employed to characterize the damage evolution during uniaxial compression. In addition, a freeze-thaw-load coupled damage constitutive model was established based on Lemaitre's equivalent strain theory and a modified Weibull distribution, and the freeze-thaw damage mechanism was further analyzed from the perspective of the interfacial transition zone (ITZ) microstructure. Results show that freeze-thaw cycling significantly reduces the compressive strength and dynamic elastic modulus of the specimens, whereas increasing the content of the all-solid-waste cementitious binder effectively enhances their resistance to freeze-thaw damage; after five freeze-thaw cycles, the compressive strength ratio increases from 45% to 79%. Owing to the structural characteristic of “strong aggregate-weak bonding,” the dynamic elastic modulus does not fully synchronize with the change in compressive strength, and is therefore more suitable as a qualitative indicator of the overall variation in internal damage. Freeze-thaw cycling also markedly alters the evolution of AE parameters during compression; in particular, after freeze-thaw cycles, specimens with a high binder content still exhibit a pronounced decrease in the b-value and continuous growth of cumulative AE energy. The proposed freeze-thaw-load coupled constitutive model can decompose the total damage into initial damage and load-induced damage. The initial damage caused by freeze-thaw action is about 29% for specimens with a binder content of 6%, whereas it is generally about 60% for the low-binder-content groups. Microscopic analysis indicates that active phases in the local Ca-Si-rich phases on the steel slag aggregate surface may participate in hydration reactions near the interface, thereby forming an ITZ denser than the mortar region; freeze-thaw cycling alters the pore-structure distribution in both the mortar region and the ITZ, damages the originally dense interfacial structure, and consequently reduces the strength and elastic modulus of the specimens.

To achieve resource disposal and utilization of lithium slag (LS), LS, ground granulated blast-furnace slag (GGBS), coal gangue (CG), calcium carbide slag (CCS), and desulfurization gypsum (DG) were used to construct an all-solid-waste cementitious system. Lithium slag lightweight aggregates (LS-LA) were prepared by cold-bonding pelletization. Four LS contents of 30%, 50%, 70%, and 90% were designed. The physical properties, single-particle compressive strength, long-term water stability and heavy metal leaching of LS-LA were tested, and its economic performance was analyzed. X-ray diffraction (XRD), scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR), thermogravimetry-derivative thermogravimetry (TG-DTG), and mercury intrusion porosimetry (MIP) were used to analyze the hydration products and pore structure evolution. The 7 d unconfined compressive strength of lightweight aggregate-hydraulically bound mixture (LA-HBM) was evaluated based on Talbot gradation. The results show that the loose bulk density of LS-LA is 946-1 021 kg·m^-3, and the apparent density is approximately 1 600 kg·m^-3, indicating a lightweight characteristic. When the LS content increases from 30% to 90%, the density and single-particle compressive strength of LS-LA decrease, whereas the water absorption and immersion mass loss rate increase. After continuous immersion for 120 d, the mass loss rates of LS-30 and LS-50 are both lower than 5%. The leaching test results show that cold-bonding pelletization reduces the release risk of heavy metals. The Mn leaching concentration decreases from 0.827 mg·L^-1 in raw LS to a minimum of 0.002 mg·L^-1 in LS-LA. Microstructural analysis shows that calcium (alumino)silicate hydrate (C—(A)—S—H) gel and ettringite (AFt) fill the particle pores and refine the pore structure. The chemical binding of gel phases and the physical encapsulation of the particle structure jointly limit heavy metal migration. The 7 d unconfined compressive strength of LA-HBM is significantly affected by gradation. The mixture with a Talbot exponent of n=0.5 shows the highest strength, but it is lower than that of the natural aggregate control group. The economic analysis shows that the material cost of LS-LA is 18.56-67.76 CNY·t^-1, and LS-50 has the lowest cost per unit strength. In summary, cold-bonding pelletization based on an all-solid-waste cementitious system can convert LS and various solid wastes into lightweight artificial aggregates with low leaching risk. LS-50 shows better comprehensive performance and economic performance, and can be used as an alternative aggregate for base or subbase layers in low-grade highways.

The lack of interfacial adhesion between iron tailings sand and asphalt restricts the long-term and durable development of iron tailings in asphalt pavement. In this paper, the fatigue characteristics of iron tailings sand asphalt mixture are studied comprehensively from the perspective of interface. Firstly, the complex modulus of iron tailings asphalt mortar was measured by frequency scanning test, and the interface interaction between iron tailings-asphalt and iron tailings asphalt mortar-fine aggregate was explained. The interface adhesion strength of iron tailings asphalt mortar-coarse aggregate was further characterized by macroscopic pull-out test. Then, the interaction mechanism of iron tailings-asphalt interface and the adhesion enhancement mechanism of cement and silane coupling agent to iron tailings asphalt mortar were revealed by infrared spectroscopy (FT-IR) and atomic force microscopy (AFM). Then, based on the linear amplitude scanning test (LAS) and semi-circular bending fatigue test (SCB), the fatigue characteristics of asphalt mortar and asphalt mixture mixed with iron tailings sand were investigated respectively, and the correlation between fatigue parameters and interface parameters of materials at each scale was constructed. The results show that the interface interaction and adhesion strength of iron tailings asphalt mortar-fine aggregate interface after cement and silane coupling agent modification are significantly improved, and the interface interaction index of iron tailings-asphalt interface after silane coupling agent modification is significantly slowed down. The effect of cement on the interfacial adhesion strength of iron tailings asphalt mortar prepared by 70 # asphalt is better, and the fatigue life of the mixture is increased by up to 80%. The silane coupling agent KH-550 has a better performance improvement effect on the iron tailings sand asphalt mixture prepared by SBS modified asphalt, and its fatigue life is the highest and increased by 82%. Through the analysis of infrared spectrum characteristic peaks, it is found that the adhesion mechanism between iron tailings and asphalt interface is mainly physical adsorption. The surface micro-roughness, micro-modulus and adhesion of iron tailings asphalt mortar modified by cement, mineral powder and silane coupling agent were significantly improved through atomic AFM. The correlation between the fatigue life of asphalt mixture and the interfacial interaction index of iron tailings asphalt mortar, interfacial adhesion strength and the fatigue life of mortar is more than 0.9. The research results are of great significance to improve the durability of solid waste-based asphalt mixture.

The performance degradation of high-RAP-content recycled asphalt mixtures due to insufficient blending between aged and virgin asphalts has become a critical constraint on their widespread application. In this study, a degree of blending (D_b) index was proposed at the mortar scale. Dynamic shear rheometer tests on asphalt binders, asphalt mortar fatigue tests, and bending beam rheometer tests were conducted to evaluate the multi-temperature-range viscoelastic properties, fatigue life, and low-temperature cracking resistance of recycled asphalts and mortars with varying RAP contents and bio-based rejuvenator conditions, thereby elucidating the performance variation patterns associated with blending states at the mortar scale. The results indicate that insufficient blending significantly increases the modulus of recycled asphalt and mortar, deteriorates rheological performance, induces pronounced strain sensitivity in fatigue behavior, and causes a sharp reduction in fatigue life under large strain levels, accompanied by a concurrent decline in low-temperature cracking resistance-effects that are more pronounced at the mortar scale. D_b values calculated from master curve shifting were 75.2% for 25% RAP mortar and 38.7% for 50% RAP mortar, while the complex modulus increment of the 50% RAP mortar was only one-sixth of that at the binder scale, confirming the significant mitigating effect of the high-modulus fine aggregate skeleton on binder-level differences. The addition of a bio-based rejuvenator markedly enhanced the blending degree, enabling the fatigue life and low-temperature cracking resistance of the 50% RAP mortar to recover to levels close to those of the 25% RAP mortar. The proposed D_b index exhibited a significant positive correlation with mortar fatigue life and low-temperature cracking resistance, effectively characterizing the interfacial blending state and providing a scientific basis for performance evaluation and design optimization of high-RAP-content recycled asphalt mixtures.

Cement-treated recycled aggregate material (CRAM), derived from construction and demolition waste, is highly susceptible to cracking during service, which significantly undermines the integrity and stability of pavement structural layers. Microbially induced calcite precipitation (MICP) has emerged as a promising technique to enhance the durability of permeable CRAM. However, its effectiveness remains insufficiently validated. This study proposed a novel crack-healing characterization method based on Peridynamics (PD) to evaluate the mechanical recovery induced by MICP. Cement mortar treated with MICP was used as the model material, and numerical simulations were performed to evaluate the flexural strength of cement mortar incorporating crack-healing behavior, as well as the flexural strength of CRAM utilizing the same type of MICP-treated mortar. The results show that the proposed PD approach accurately captures the recovery of mechanical performance due to crack healing. While MICP treatment significantly restores the flexural strength of cement mortar, its effect on the flexural strength of permeable CRAM is limited. Moreover, reduced cement content markedly diminishes the healing efficiency; when the matrix filling ratio falls below 0.76, strength recovery becomes negligible even after 28 days of curing. Extending the healing-related curing time effectively enhances tensile strength recovery while mitigating crack branching. A minimum of 7 days of healing is recommended to ensure meaningful mechanical restoration through MICP.

To investigate the damage evolution mechanism of low-temperature cracking in warm-mix recycled asphalt mixtures under freeze-thaw (FT) cycles, low-temperature bending tests were conducted on a conventional asphalt mixture (AM) and two types of warm-mix recycled asphalt mixtures (WAM-S and WAM-M), with simultaneous acoustic emission monitoring. The results showed that, after 15 FT cycles, the flexural tensile strength and maximum flexural tensile strain of all three mixtures decreased by more than 30%. Crack evolution gradually shifted from small-scale propagation to dominance by large-scale cracks, and the crack morphology became dominated by tensile cracks, with their proportion exceeding 90%. In addition, FT cycles reduced the contribution of mixed failure and aggregate failure during the stages of microcrack development and crack propagation to final failure, causing the damage mechanism to evolve from the parallel development of multiple failure modes to the dominance of cohesive and adhesive failure. A comparison between the two warm-mix recycled asphalt mixtures indicated that WAM-S exhibited larger crack scales and more pronounced dominance of tensile cracks and adhesive failure, whereas WAM-M was able to delay crack scale growth and crack morphology convergence, demonstrating superior freeze-thaw resistance. Correlation analysis further confirmed that the proportion of tensile cracks was significantly negatively correlated with low-temperature crack resistance, whereas the relationships between different damage mechanism clusters and low-temperature crack resistance exhibited clear stage-dependent characteristics. Therefore, delaying crack concentration and crack morphology convergence through interfacial regulation is an important strategy for improving the low-temperature crack resistance of warm-mix recycled asphalt mixtures.

To promote the resource-based and large-scale utilization of solid waste materials such as silt in subgrade engineering, an inorganic composite curing agent and fluidized solidified silt (FSS) were prepared using slag (SL), high belite sulphoaluminate cement (HBSC), gypsum (G), and calcium carbide residue (CCR). The optimal mixing ratio of inorganic composite curing agent was recommended. The effect of mud moisture content and curing agent content on mechanical properties and deformation characteristics of FSS were explored. The strengthening mechanism of FSS was also revealed. The results showed that the mass variation of CCR has the most significant impact on the properties of FSS. The optimal mass ratio of inorganic composite curing agent was m_SL∶m_HBSC∶m_G∶m_CCR=1∶0.8∶0.6∶0.2 for selected coastal silt. Considering the mechanical strength and deformation characteristics, the recommended range for mud moisture content is 79% to 85%, and the curing agent content is 12% to 16%. The 7 d unconfined compressive strength (UCS) of FSS prepared according to the recommended proportion is higher than 0.6 MPa. The dry shrinkage rate and autogenous shrinkage rate are lower than 2.5% and 2%, respectively. There is a high correlation between the peak strain, deformation modulus, and UCS of FSS, which follows power function models and linear models, respectively. After the composition of curing agent is determined, the correlations among the properties such as fluidity, setting time, bleeding rate, and UCS increase significantly. The hydration reaction, alkali activated reaction, and ion-exchange reaction of inorganic composite curing agent jointly promote the formation strength of FSS. The fluidized strengthening of silt with inorganic composite materials is beneficial for reducing cement consumption, and improving early strength, fully meeting the bearing requirements of subgrade engineering construction.

Structural Health Monitoring (SHM) technology has become a key tool for ensuring the operational safety of bridges and supporting maintenance decision-making. However, the conventional bridge SHM paradigm relies on high-cost commercial sensors, closed data acquisition devices, and centralized analysis, resulting in high costs, limited scalability, and data transmission burdens. In this context, edge intelligence, with on-site information extraction and rapid response as its core, provides a new system-level pathway for bridge SHM to alleviate constraints related to cost, data transmission, and scalability. Low-cost sensors and embedded platforms respectively constitute the front-end sensing foundation and edge computing carrier of this paradigm, jointly promoting the transition of bridge SHM from the conventional paradigm characterized by high-cost device dependence, centralized raw data uploading, and cloud-diagnosis-oriented analysis toward a lightweight edge-intelligent paradigm characterized by low-cost dense sensing, scalable and controllable deployment, and edge-side information extraction. Following this main thread, this paper systematically reviews edge-intelligence-based lightweight bridge health monitoring. Firstly, the research status and paradigm evolution of lightweight bridge SHM are summarized. Secondly, from the sensing layer, the applications and technical boundaries of three types of low-cost sensors in bridge SHM are reviewed, including inertial and mechanical sensors, environmental and acoustic sensors, and optical imaging sensors. The common error sources and error calibration strategies of low-cost sensors are also summarized. Thirdly, from the transmission and analysis layer, the hardware performance of embedded platforms is systematically organized, based on which an in-depth analysis is provided of three typical architectures in lightweight bridge SHM systems: edge acquisition nodes, edge gateway nodes, and edge computing nodes. Subsequently, the system characteristics of lightweight bridge SHM is comprehensively reviewed from three dimensions: communication mechanisms, deployment architectures, and application scenarios. Finally, current technical bottlenecks are summarized, and future development trends are discussed in terms of multimodal sensing, edge-adaptive intelligence, and intelligent maintenance decision-making.

To address high system complexity, prohibitive costs, and scalability challenges in monitoring short- and medium-span bridges, this study proposes a lightweight monitoring system and evaluation framework based on a single deflection indicator. First, a hierarchical architecture integrating “precise perception, intelligent processing, and dynamic evaluation” is established using deflection as the core monitoring indicator. Second, a signal decomposition method combining variational mode decomposition (VMD) and low-pass filtering is introduced to decouple temperature effects and high-frequency noise, thereby extracting live load-induced deflection components. Simultaneously, an assessment model is developed for structural equivalent load back-calculation and stiffness evaluation based on real-time deflection data, complemented by a three-level (blue, yellow, and red) early warning threshold scheme. Furthermore, a long-term performance evaluation and service life prediction model is constructed by incorporating reliability theory and extreme value statistics. Field validation on an actual bridge demonstrates that the system ensures stable data acquisition and accurate load effect separation, with early warning responses aligning with structural behavior and long-term reliability indices satisfying regulatory standards. The results indicate that the proposed method significantly reduces system cost and complexity without compromising monitoring accuracy, providing an effective solution for the economical and intelligent monitoring of short- and medium-span bridges.

To enable real-time online monitoring of the service condition of bridge substructures under unknown operational loads, a lightweight monitoring method for substructures based on the joint load-parameter-response estimation and sparse observation responses is proposed. Firstly, a modified adaptive unscented Kalman filter with unknown input algorithm was developed, integrated with a dynamic analysis model of the bridge substructure to construct a joint load-parameter-response estimation framework suitable for real-time online assessment of substructures. Subsequently, numerical simulations were conducted to validate the effectiveness of the algorithm under noise interference and random load conditions, and the impact of initial structural parameter errors on the robustness of the algorithm was investigated. Finally, the feasibility and effectiveness of the proposed method in complex operational environments were verified using field monitoring data from an actual bridge substructure. The results demonstrate that, when only acceleration responses at the pier top and bottom are observed, the proposed algorithm can effectively achieve simultaneous estimation of unknown loads, structural parameters, and responses of the bridge substructure, with identified structural responses and unknown random loads showing high consistency with true values, and the drift phenomena caused by noise interference and cumulative errors is effectively suppressed. Under a 5% noise level and 50% initial structural parameter error, the identification errors for structural parameters, responses, and unknown loads are all within 5%, and the pier foundation stiffness and pier body stiffness converge accurately and rapidly to their true values. Field monitoring tests further indicate that the method can effectively identify the structural parameters and dynamic characteristics of bridge substructures under unknown operational loads, with the identified time-frequency domain responses exhibiting good agreement with measured data.

To address the problems of high computational complexity and low efficiency associated with conventional evaluation methods for small-and medium-span bridges under special heavy-load passage scenarios, this study aims to develop a lightweight assessment approach to improve the efficiency of traffic safety evaluation for such bridges. A hybrid modeling framework integrating mechanical mechanisms and data-driven techniques is adopted, following a progressive workflow of “theoretical modeling-data correction-engineering validation”. First, based on the traveling characteristics of special heavy loads, a parametric displacement influence line formula was derived using the three-moment equation. A displacement amplification factor K_amp was introduced to modify the transverse load distribution coefficient, thereby establishing a practical theoretical method for rapid displacement estimation. Second, using field monitoring data from 40 bridges, a displacement error correction model were constructed by incorporating machine learning techniques. Finally, case studies are conducted on test bridges, and the results were compared with those obtained from finite element analysis for validation. The results show that the theoretical displacement extraction method, verified by field measurements, can cover the actual displacement with a partial safety margin (coverage range of 26%), meeting the engineering accuracy requirements. The displacement error correction model reduces the absolute error from 0-7 mm to 0-3 mm, and the mean absolute error (MAE) decreases from 3.351 mm to 1.778 mm, with an average error reduction rate of 17.06%. The goodness of fit between theoretical and measured data (R²) increases from 0.308 to 0.752. SHAP interpretability analysis shows that stiffness (EI) contributes the most to the error model output (37.36%, positively correlated), stiffness ratio (R_k) shows a stable negative correlation, and the displacement lateral distribution coefficient (η) has a weak influence. Case comparison results demonstrate that the proposed method shows good agreement with finite element analysis results and significantly improves computational efficiency while maintaining evaluation accuracy. By deeply coupling mechanical mechanisms with data-driven modeling, the proposed lightweight assessment approach effectively overcomes the limitations of traditional methods and provides a reliable technical pathway for enhancing traffic safety evaluation of small-and medium-span bridges under special heavy-load conditions.

Waveguide rods can be used for acoustic emission (AE) monitoring of broken strands in prestressed steel strands of bridges. However, the attenuation characteristics of AE waves propagating in waveguide rods result in short propagation distances and limited monitoring ranges. Consequently, a large number of sensors must be deployed to meet monitoring requirements, increasing the economic cost of monitoring. To expand the monitoring range, firstly, a theoretical model for AE waves propagation in acoustic black hole (ABH) waveguide rod was established, and a geometric model of a cylindrical waveguide incorporating an ABH structure was constructed. Then, the propagation mechanism of AE waves in the ABH waveguide was visualized through numerical simulation methods. The effects of ABH length and power exponent on the signal amplification efficiency of damage source AE waves are analyzed. Finally, through full-scale high-speed railway box girder engineering tests, the feasibility of the ABH waveguide rod structure for engineering applications was validated. Results demonstrate that the proposed ABH waveguide rod structure effectively amplifies AE signals. AE waves undergo counterclockwise reflection within truncated ABH, with wave amplitudes increasing according to a power-law function as propagation approaches the truncation height. Both ABH length and power exponent exhibit proportional relationships with signal amplification efficiency. Therefore, the proposed ABH waveguide rod wave-guiding method achieves signal amplification, effectively extends propagation distance, enhances monitoring coverage, and reduces monitoring costs. The study provides a theoretical foundation and engineering application reference for implementing the ABH waveguide rod wave-guiding method in practical structural health monitoring.

To enable bridge weigh-in-motion (B-WIM) under realistic vehicle driving conditions, this paper proposed a novel method integrating computer vision and the Fast Dynamic Time Warping (FastDTW) algorithm. The approach eliminated traditional assumptions of constant-speed straight-line travel while maintaining accuracy even with uncertain bridge entry/exit timings. In the specific operation process, computer vision technology was first used to identify and track the vehicle to obtain its spatio-temporal information. Then, combined with the vehicle weight statistics, the vehicle load was initialized to provide raw data for the subsequent iterative solving of the real vehicle load. Next, with the help of the FastDTW algorithm, the bridge response was associated with the measured time curve, and the axle distance and axle weight information were adjusted through continuous iteration, so that the two were gradually approximated to obtain the accurate axle distance and axle weight of the vehicle. After simulation and model test verification, the average relative error of axle distance and axle weight identification was only 2.83% and 2.67%, the maximum relative error was 5.60% and 5.84%, and the maximum relative error of total weight was 4.28%. Moreover, in the validation with real bridge tests, the relative errors in identifying the maximum axle distance and maximum axle weight were 10.03% and 6.16%, respectively, while the relative error in identifying the maximum total weight was 5.72%. These results fully demonstrate the accuracy and applicability of the proposed algorithm, indicating its broad potential for practical applications.

Hinge joints are key components for lateral connection and internal force transfer in hollow slab girder bridges. During long-term service, hinge joint deterioration weakens the lateral cooperative mechanical performance of the bridge, increases the risk of single-slab action under vehicle loads, and affects bearing capacity and service safety. In current engineering practice, the deterioration of load-transfer performance in hollow slab girder bridges is still mainly evaluated by visual inspection of hinge joints. Such evaluation is largely qualitative and subjective, and lacks an objective characterization of the deterioration degree of transverse load-transfer performance. To address this issue, this study proposes a rapid load testing method for fast screening of transverse load-transfer performance deterioration in hollow slab girder bridges. First, the relationship between hinge joint stiffness and slab deflection and strain responses is derived, and the transverse response redistribution caused by hinge joint damage is clarified. On this basis, a rapid moving-load testing procedure is established, and a response relative-difference index together with its local mutation criterion is constructed to realize rapid screening, localization, and comparative evaluation of hinge joint damage. Furthermore, based on the qualitative identification results, a model-updating-based quantitative evaluation method is introduced. An objective function based on the transverse response distribution confidence criterion is established to quantitatively evaluate the deterioration degree of transverse load-transfer performance of damaged hinge joints. Finally, the proposed method is verified on an urban viaduct hollow slab girder bridge with a span of 22 m and 16 slab girders. The field test included seven loading cases and lasted approximately 30 min. The identification results could effectively reflect the location characteristics of the main damaged hinge joints, and were generally consistent with the main disease regions revealed by on-site visual inspection. The results show that the proposed method can provide technical support for rapid screening and quantitative evaluation of load-transfer performance deterioration in existing hollow slab girder bridges, and has practical engineering application value.

To address the issue of low accuracy in existing short-term wind speed prediction methods, this paper proposed a hybrid forecasting model based on secondary decomposition using sample entropy. The original wind speed series were first decomposed using Empirical Mode Decomposition (EMD), and the complexity of each resulting sub-series were quantified through the calculation of its sample entropy. Subsequently, sub-series with higher complexity were further decomposed by means of Variational Mode Decomposition (VMD) as a secondary processing step. The decomposed components were then reconstructed using the K-means method, whereby sub-sequences were grouped based on similar entropy values. Finally, a combined forecasting model integrating a Gated Recurrent Unit (GRU) network and a Bidirectional Long Short-Term Memory (BiLSTM) network were developed for each reconstructed sub-series according to its complexity level. The overall forecast was obtained through the aggregation of the predictions from all sub-series. Validation with measured wind farm data demonstrates that the Mean Absolute Error (MAE) is reduced by more than 50% compared to common single prediction models. A reduction in MAE of over 25% is also achieved when compared to both low-hybrid models with single decomposition and dual forecasting models with secondary decomposition. The effectiveness and stability of the proposed model in short-term wind speed prediction are thus sufficiently verified, while also exhibiting strong generalization capability.

Prestressed concrete (PC) bridges are widely used in the road network and often suffer structural damage such as cracks during service. Timely and effective assessment of these damages is crucial for ensuring bridge operation safety. Cracks serve as a visual manifestation of structural damage. Utilizing digital image technology for multi-faceted characterization of cracks and establishing their intrinsic relationship with structural damage states represents an effective approach for rapidly evaluating structural conditions. Accordingly, a multi-feature characterization method integrating explicit geometric texture features and implicit multifractal features of crack images was proposed in this paper, constructing a feature system with forty-five quantitative indicators. Through static failure tests of PC beams with different corrosion levels, images of the entire crack propagation process were systematically collected. Combined with the load-bearing ratio (as a damage factor), a damage prediction dataset containing 113 valid samples was established. Using various machine learning algorithms, this study systematically compared modeling performance under three feature engineering strategies: combined features, explicit features only, and multifractal features only. The consistency of model performance across different test beams and the generalization capability across structures were evaluated. Results show that features extracted from crack images correlate clearly with the damage factor, with explicit features exhibiting more pronounced correlations. Multifractal spectrum parameters and generalized fractal dimensions evolve regularly with increasing load, showing distinct distribution characteristics in intervals of positive and negative weight coefficients, which validates the effectiveness of the proposed multi-scale crack quantification system. Across all strategies, the MLP model achieves superior robustness in modeling the damage factor. The combined feature strategy achieves significantly better prediction accuracy than any single-feature strategy. Models show strong consistency across different test beams, with minor influence from individual sample variations. In structural-level transfer experiments, the combined feature strategy exhibits the best generalization ability, significantly outperforming the explicit-feature-only strategy, indicating that multifractal features provide valuable supplementary information in cross-structure prediction. The proposed explicit-implicit multi-scale feature fusion framework effectively reveals the intrinsic relationship between crack evolution and structural damage, offering a reliable approach for intelligent crack identification and quantitative damage assessment of highway bridges.

Traditional health monitoring systems designed for large-scale bridges are often costly, complex to deploy, and dependent on power and communication infrastructure, making them unsuitable for widespread deployment on the vast number of small- and medium-span bridges. With recent advances in AI and image processing technologies, vision-based bridge health monitoring has emerged as a competitive alternative to traditional monitoring approaches, owing to its non-contact nature, light weight, and flexible deployment. This paper focuses on vision-based bridge vibration monitoring, following a “perception, identification, deployment” framework. Based on a comprehensive review of existing literature, three key research tasks are systematically undertaken to address the critical challenges across these stages. Firstly, at the perception stage, automated 2D/3D displacement measurement methods are proposed, combined with coded target design and image processing algorithms, to realize the efficient perception of multi-target, multi-plane and multi-dimensional displacement of bridges, and the measurement accuracy and stability are verified by the bridge model shaking table test. Secondly, at the identification stage, a vision-based moving vehicle load identification method is constructed to realize the technical leap from “response monitoring” to “load identification”, and its effectiveness is verified through the vehicle-bridge dynamic interaction experiments. Finally, at the deployment stage, an embedded vision sensing system integrating Internet of Things and edge computing is developed, which can obtain the vibration characteristics of cables and estimate the tension in real-time, effectively reducing the latency caused by large-scale data transmission and offline processing. In summary, this study presents substantial contributions in both review and method development, providing a new path for intelligent, lightweight and scalable bridge health monitoring, demonstrating strong adaptability and engineering potential.

Traditional time-varying performance evaluation of bridges mainly relies on static or low-speed load tests, which require traffic interruption, incur high operational costs, and offer limited efficiency. Moreover, these discrete test results are highly sensitive to environmental and traffic variability, making it difficult to capture the continuous evolution of structural performance. To address these limitations, this study proposes a quantitative evaluation framework for bridge time-varying performance based on responses induced by random traffic flow, and develops a complete methodology encompassing theoretical modeling, indicator derivation, and field validation. The main contributions are as follows: ① a random traffic flow model was constructed based on probabilistic theory, and its distributional characteristics were analyzed through numerical simulations and monitoring data from the Wuhu Yangtze River Second Bridge; ② by integrating a hierarchical Bayesian data fusion framework with the ensemble empirical mode decomposition (EEMD) theory, a new time-varying performance evaluation method for bridges under random traffic-induced deflection monitoring was developed, and a quantitative relationship between stiffness degradation and the residual deviation rate was analytically derived; and ③ field validation was conducted using long-term monitoring data from the Shibu Bridge in Wenzhou, enabling assessment of structural performance before and after reinforcement. Results demonstrate that the proposed method exhibits strong adaptability and robustness across various random traffic flow conditions. Field measurements further show that the residual deviation rates of the dynamic deflection responses decreased by 24.57% and 17.68% after reinforcement, corresponding to an approximate 10%-20% increase in global bridge stiffness, thereby confirming the feasibility of monitoring structural performance evolution without interrupting traffic.

To advance technical approaches for prestressed structure detection and intelligent assessment, this research addressed theoretical gaps and standardization issues in anchorage prestress evaluation. Focusing on effective prestress assessment for post-tensioned precast girders, the study employed field measurements from an in-construction highway bridge. Three clustering algorithms were developed using the elbow method, silhouette coefficient, and parameter combinations to establish the characteristic value for effective anchorage prestress. Five-fold cross-validation with L1 and L2 regularization was implemented to mitigate overfitting, while grid search optimized four regression models to define the strand bundle's mean force tolerance range. Results confirm K-means and hierarchical clustering as optimal based on clustering metrics and engineering practicality, where effective anchorage prestress characteristic values are defined as: When the tensioning coefficient is 0.75, for tendon lengths not exceeding 30 m, the standard value is 178 kN; for lengths exceeding 30 m, the standard value is 180 kN. When the tensioning coefficient is 0.7 and tendon length exceeds 40 m, the standard value is 171 kN. The XGBoost model, optimized through evaluation metrics, integrates conformal prediction to deliver prestress prediction intervals across confidence levels, establishing ±6 kN as the acceptable tolerance range for effective prestress along full strand bundles.

Drilling and blasting tunnel construction equipment configuration has poor adaptability, and equipment scheduling and management are challenging in the complex underground environment. To ensure the safe operation and efficient passage of tunnel construction equipment, and to enhance the mechanization, intelligence, and efficiency of the entire drilling and blasting tunnel construction process, this study conducts research and application on the dynamic configuration and intelligent scheduling system of construction equipment. By analyzing the passage modes and road characteristics of drilling and blasting tunnel construction equipment, an improved multi-machine scheduling algorithm based on heuristic function optimization, search optimization, and modeling optimization is proposed. A UWB and TOF fusion-based two-dimensional precise positioning method for tunnels and a dynamic path planning method for equipment optimization are established. MATLAB simulation results demonstrate a 22.5% improvement in path search efficiency and obstacle avoidance route planning time reduced to less than 1 s, validating the feasibility and efficiency of this optimization algorithm for tunnel equipment path dynamic planning. Furthermore, a time-cost multi-objective optimization model for critical equipment configuration is established using the NSGA-Ⅱ algorithm and entropy-weighted TOPSIS decision-making, this forms a comprehensive equipment configuration method for all construction processes, considering tunnel cross-section types, rock mass conditions, excavation mileage, and construction schedule requirements. An intelligent scheduling system for drilling and blasting tunnel is developed, integrating dynamic equipment configuration, real-time dispatch, simulated scheduling, and construction planning. This system has been applied to a railway tunnel project. Application results indicate: 38.9% reduction in process transition time, 19.7% decrease in construction operation time, and a 21.5% overall improvement in construction efficiency, validating the system's significant application value for drill-and-blast tunnel projects.

To promote the high-value utilization of tunnel muck aggregates in concrete and clarify the effects of aggregate lithology on concrete performance, tunnel muck aggregates of four lithologies, namely limestone, gneiss, sandstone, and granite, were selected in this study. The effects of aggregate lithology on the workability, mechanical properties, and steel-concrete bond performance of concrete with different strength grades were investigated, and a compressive strength prediction model was established. The results show that, to achieve the same slump, the required dosage of water-reducing admixture increases in the order: limestonesandstone>gneiss>granite. A modified Abrams water-binder ratio-strength model considering the effects of stone powder content and methylene blue (MB) value was established, and it can effectively predict the 28-day compressive strength. Microscopic analysis shows that aggregate lithology affects the continuity of hydration products and the compactness of the interfacial transition zone in concrete. The filling and nucleation effects of limestone powder help improve the compactness of the matrix and interfacial transition zone, whereas the high stone powder content, high MB value, and low stone powder activity of granite manufactured sand lead to increased pores and interfacial defects, thereby weakening the stress transfer capacity. The porosity of the matrix and interfacial transition zone increases in the order: limestone

With the rapid advancement of connected autonomous vehicle (CAV) technology, CAVs leverage precise environmental perception and autonomous control capabilities to progressively eliminate safety hazards caused by human errors. This enables safe and efficient operation in a lane-free environment, significantly improving the utilization of lateral road space, which remains largely underutilized in traditional lane-based human-driven traffic. First, focuses on lane-free traffic of CAVs on freeways and proposes a novel traffic control strategy “Internal Boundary Control” (IBC). Using this control method, dynamic adaptation between bidirectional road resources and traffic demand is achieved. Additionally, for scenarios where the total bidirectional traffic demand (including on-ramp flows) exceeds the total road capacity, the paper systematically explores a coordinated control mechanism for internal boundaries and on-ramp flows to achieve spatiotemporal optimization of road traffic efficiency in both directions. Then, a traffic flow modeling framework tailored for lane-free traffic of CAVs on freeways is developed, featuring an internal boundary-adaptive fundamental diagram model and an extension of the classical first-order cell transmission model. Finally， based on this model, the coordinated control of internal boundaries and on-ramp flows is formulated as a quadratic programming model with the objective of minimizing total travel time. The results show that the developed macroscopic model accurately captures the dynamic evolution characteristics of lane-free traffic flow and effectively simulates critical phenomena such as capacity drop at bottleneck sections. Both IBC and the coordinated control strategy significantly enhance system operational efficiency, compared to the no-control scenario, IBC alone reduces network-wide travel delay by approximately 57%, while the coordinated control achieves a reduction of 78%. The lane-free traffic paradigm of CAVs represents a fundamental shift in transportation systems, while IBC provides an innovative mechanism for the elastic allocation of road resources. This study offers important theoretical and methodological support for the design and performance optimization of future management and control systems for CAV traffic on freeways.

Modular bus technology allows different buses to form bus platoon by speed coordination and physical formation during operation, and allows passengers to in-motion transfer without stopping in the platoon. To explore the potential advantages of modular feeder bus technology in demand-responsive services and alleviate the 'first and last mile’ of the real problem, this paper designed a flexible scheduling system for modular feeder buses, constructed a mixed integer non-linear programming (MINLP) model based on the spatial-temporal network modeling method with the objective of minimizing the buses'operational cost and the passengers’ travel time cost, which can realize the optimization of modular bus routes, and the passenger-to-modular bus assignment, bus formation plan and in-motion plan simultaneously under reservation travel demand. For the feature of the constructed model, the linearization was developed to reconstruct the model into a mixed integer linear programming (MILP) model and a specific search space reduction strategy was designed to improve the solution efficiency. In order to verify the effectiveness of the constructed model, numerical experiments were conducted based on a real case in some areas of Shijingshan District, Beijing, and sensitivity analysis was performed. The experimental results show that compared with the conventional demand-responsive bus system, the modular bus flexible scheduling system can reduce the total system cost by about 5.50%-10.94% and the operational cost by about 16.60%-23.36% in cases of different time periods, demonstrating significant advantages in reducing operational costs and effectively improving the comprehensive benefits of demand-responsive feeder services. The sensitivity analysis results can provide a decision basis for the design of modular bus capacity and the number of buses to be invested in actual operation.

Full-vector wire-controlled multi-axle heavy-duty vehicles with high load, high mobility and high throughput are important transport equipment for national defence and material delivery. Aiming at the difficult problem that stability and manoeuvrability are difficult to be taken care of at the same time under the extreme working conditions of full vector four-axis unmanned heavy-duty vehicles (8WID-8WIS), this paper proposes a transverse and longitudinal hierarchical collaborative control strategy based on motion tracking for full vector four-axis unmanned heavy-duty vehicles. First, based on the optimal control method for Zero Side-slip Angle, an adaptive LQR trajectory tracking model is established to calculate the desired driving state under all-wheel steering. Second, a three-layer chassis domain integrated control architecture is established by designing the horizontal and vertical layered cooperative control strategy: the top layer reference is the state generator, responsible for solving the global generalized force corresponding to the desired driving state; the middle layer is the tyre force allocation layer, which achieves optimal allocation of tyre forces at the tyre layer; the bottom layer is the drive execution layer, which combines the Dugoff tyre inverse model to complete the dynamic mapping of tyre forces to the drive space, thereby achieving coordinated control of path tracking and force allocation for unmanned heavy-duty vehicles. Finally, the proposed method is verified based on the joint simulation platform of MATLAB/Simulink and TruckSim. The results show that the proposed a transverse and longitudinal hierarchical collaborative control strategy can significantly improve the extreme motion stability of multi-axle heavy-duty vehicles, more accurately track the desired motion state and expand the stability margin, while effectively balancing the global dynamic performance of the multi-axle vehicle chassis, providing reference value for the integrated control of unmanned heavy-duty vehicles.

The nonlinear friction characteristics between the upper column in the steer-by-wire (SBW) system significantly affect the control precision of the steering system. However, traditional friction models inadequately consider characteristics such as friction hysteresis and friction backlash, making it difficult to accurately describe the nonlinear friction characteristics of the upper column in the SBW system. This study focuses on the internal friction characteristics of the upper column in the SBW system and proposes an improved Bouc-Wen friction model. Firstly, based on the nonlinear friction characteristics of the upper column in the SBW system, a comprehensive analysis is conducted by combining the Stribeck friction effect and the friction hysteresis characteristics. Secondly, by introducing parameters related to the acceleration-dependent friction force and the friction overshoot characteristics into the normalized Bouc-Wen model combined with the Stribeck effect, an improved Bouc-Wen friction model suitable for the upper column of the SBW system is proposed. Then, a combination of bench testing and the Particle Swarm Optimization (PSO) algorithm is employed to accurately identify key parameters such as M_e, δ, ρ and σ in the improved Bouc-Wen model. The experimental data indicate that the improved Bouc-Wen model has a significant advantage in describing the frictional hysteresis and hysteresis characteristics of the column, with the predicted friction hysteresis loop fitting well with the measured data, showing a more plump closed curve loop, thus verifying the effectiveness of the model. Finally, the improved model outperforms traditional models in friction compensation prediction, with a Root Mean Square Error (RMSE) of 0.133 5 and a Mean Absolute Error (MAE) of 0.084 1. It accurately captures the characteristics of friction data and efficiently identifies parameters through the PSO algorithm, effectively characterizing the internal friction of the column in the SBW system. This study provides a new theoretical and methodological approach for friction compensation of the column in the SBW system, laying a foundation for subsequent research on the simulation of steering feel torque by the road feel motor.