Thermosetting polyurethane-modified asphalt (PUMA) undergoes an irreversible curing reaction, which means any improper handling during construction cannot be corrected. This study investigates the effects of hard segment content, isocyanate index, polyurethane (PU) content, and temperature on the viscosity of PUMA and proposes a method for regulating its reserve time. Based on viscosity-temperature profiles, a viscosity growth model was established to predict the evolution of PUMA viscosity with temperature and time. The results indicate that temperature exerts the most pronounced effect on the reserve time of thermosetting PUMA. Moreover, the developed viscosity growth model under different temperatures enables the prediction of PUMA reserve time. As PU content increases, the microstructure of PUMA evolves from a dispersed to a continuous phase. When the PU content exceeds 30%, a continuous crosslinked network forms, imparting excellent thermal resistance and elastic recovery at high temperatures. Furthermore, high PU content enhances the low-temperature flexibility of PUMA. These findings provide valuable insights for optimizing the application of PUMA materials in pavement engineering.

Tunnels often traverse variable and complex strata， thereby posing significant challenges in analyzing long-term tunnel service performance. This study proposes the tunnel service performance deep surrogate model （TSP-DSM）， an intelligent prediction framework designed for predicting long-term tunnel service performance under complex geological conditions. The TSP-DSM framework employs DenseNet as a deep surrogate model， extracting multilevel feature representations from tunnel geological cross-sectional diagrams and horizontal-convergence monitoring data， and adaptively learning the potential nonlinear mapping relationship between geological conditions and service performance. To increase the prediction accuracy， the model’s hyperparameters are optimized via random grid search. Experimental results demonstrate that the TSP-DSM achieves a training accuracy of 91.29%， outperforming GoogleNet by 12.79% and ResNet by 5.72%； the prediction performance achieved using grayscale images is comparable to that attained using RGB images. On test samples， which were sourced from complex strata， the TSP-DSM achieves a prediction accuracy of 78.19%， demonstrating strong generalization across various strata. A tunnel service-performance visualization map， constructed based on the intelligent prediction results， provides an intuitive basis for tunnel condition assessment and risk prediction.

To address the challenge of distinguishing subjective aggressive driving (initiated by drivers) from hazardous behaviors caused by external cyberattacks, this study proposes an innovative intent recognition framework named Intent-Decipher. By integrating the information credibility outputted by an intrusion detection system (IDS) into a security-aware inverse reinforcement learning (SA-IRL) model, the framework infers the reward function behind vehicle behaviors and classifies three key driving intents: normal, aggressive, and malicious. Experiments were conducted on a semi-synthetic dataset containing 20 000 trajectories. Results show that Intent-Decipher significantly outperforms baseline methods in classification accuracy, achieving a macro-average F1-score of 0.94. Notably, Intent-Decipher excels at differentiating subjective aggressive driving from attack-induced behaviors: its F1-score for identifying malicious attack-induced (MAI) intent reaches 0.90, an absolute improvement of 0.16 compared with the standard inverse reinforcement learning (IRL) model (which lacks security awareness and only achieves an F1-score of 0.74).

To investigate the bearing behavior and failure modes of six‐pile thick pile caps under different reinforcement configurations and explore the optimal reinforcement scheme， this study examined four scaled specimens （S1‐S4） with distinct reinforcement designs. The bearing capacity of each pile cap was first calculated using various methods， and laboratory tests were then conducted to determine cracking and ultimate loads. Reinforcement stresses and key strain measurements in the pile caps were monitored， and the crack propagation process was documented in detail. The results demonstrate that the spatial truss model yielded calculations closest to experimental values. Specimen S3 with mesh reinforcement exhibited the highest bearing capacity but required greater steel consumption. The truss‐reinforced S4 showed enhanced ductility at failure but posed constructability challenges. Uniformly reinforced S1 delivered the lowest bearing capacity and developed more uneven cracks. Furthermore， a comprehensive analysis of reinforcement stress distribution， internal force flow transfer， and the validity of the plane‐section assumption at the pile‐cap sides revealed that the mechanical behavior of the six‐pile thick pile cap is more closely aligned with the spatial truss model. The concentrated reinforcement scheme at the pile head， as suggested by this model， proves to be an efficient and practical design solution.

The offshore rocket recovery tower is an essential infrastructure for reusable space transportation. To address the lack of design methods and quantitative performance evaluation approaches for prestressed cable-stayed bracing systems under asymmetric and complex loading, this study develops an asymmetric two-stage prestressed cable-stayed bracing system for a large-scale offshore recovery tower with a height of 67 m and plan dimensions of 54 m × 70 m. Dual-platform finite element models were established in MIDAS Gen and SAP2000, and 292 full-condition load combinations were analyzed using a nonlinear step-by-step inheritance algorithm. In parallel, a refined Abaqus solid model was employed to verify the core load-bearing joint locally. Prestressing reduced the maximum X-direction deformation under transportation conditions from 263.2 to 222.1 mm, yielding safety margins of 14.6% and 17.5% in the X- and Y-directions, respectively. The stresses in the main structural members were markedly improved—the Q460 truss and beam elements satisfied the allowable-stress limits, whereas the Q355 truss elements remained only slightly above the limit. The local maximum von Mises stress at the inner pulley block joint reached 891.9 MPa, indicating a risk of local yielding. These results demonstrate that the asymmetric prestressed cable-stayed bracing system can effectively enhance the global stiffness and force-transfer performance of large-scale offshore recovery towers, and that the full-process analytical framework established in this study provides a theoretical basis and engineering reference for the design and performance evaluation of analogous offshore aerospace support structures.

Usual environmental microvibrations (UEMs) can introduce inaccuracies for field-measured traffic-induced vibration responses, leading to significant errors in the inversion of highway traffic loads. This study proposed a method for eliminating UEMs in combination with the inverse pseudoexcitation method to invert highway traffic loads. The real vibration response caused by traffic loads was obtained by separating the UEM component from the measured ground vibration response of a site through a frequency-domain transformation. The frequency response function was subsequently derived using finite element simulation technology. Finally, the traffic load spectrum was inversely identified by performing a matrix inversion operation on the frequency response function matrix. The vibration response at a measurement point was calculated using the inverted traffic load spectrum from a synchrotron radiation light source project and compared with field measurements. The relative error decreased from 21.15% to 6.82%, validating the effectiveness of the proposed method. This approach avoided a complex vehicle-road coupling analysis and provided a reliable technical means for load inversion at sites sensitive to microvibration.

To investigate the aerodynamic performance of geometrically eccentric structures, wind tunnel testing and numerical simulations were employed to examine the flow field characteristics and tri-component aerodynamic forces of T-shaped towers with varying geometric eccentricities. A mathematical model for their generalized force spectra was established. The reliability of the numerical simulation method was validated by comparing the lift, drag, and torque coefficients obtained from wind tunnel experiments with the numerical results. Subsequently, the influence of geometric eccentricity on wind pressure distribution, vortex shedding characteristics, and turbulent kinetic energy distribution was analyzed. Using random vibration theory, a normalized formula for the full-tower force spectrum of the T-shaped tower was derived through nonlinear fitting of numerical simulation data. The results indicate that with increasing eccentricity, the vortex shedding of the structure becomes more complex, with a bias toward the eccentric side. Its wake morphology undergoes distortion and deformation, exhibiting vertical extensibility and leading to peak wind pressure occurring at the longer cross-arm. Furthermore, the torque coefficient demonstrated the highest sensitivity to eccentricity variations. A 50% increase in eccentricity resulted in a 99.7% increase in the torque coefficient.

Prestressed concrete box girders are widely used in bridge engineering. Effective monitoring to accurately detect, localize, and evaluate damage development in in-service box girders is essential to ensure their operational safety. In this study, a four-point bending test based on acoustic emission (AE) technology was conducted on a full-scale in-service box girder with 20 years of service. The correlation between damage evolution and AE parameters was investigated, laying the foundation for the application of AE technology in bridge monitoring. The results show that the cracking load of the box girder after 20 years of service is lower than the theoretical design value. Changes in the cumulative AE ring counts can characterize different failure stages of the box girder. AE signals were grouped by amplitude and peak frequency for analysis, and the effectiveness of different AE signals in tracking damage development was examined. The Gaussian mixture model (GMM) clustering algorithm was used to perform the clustering analysis of rise time/peak amplitude (RA) and ring-down counts/duration (AF). This method can effectively distinguish microcrack types and identify failure modes in different damage stages. In conclusion, AE monitoring can characterize the structural damage state and improve the safety and reliability of box girders.

With the development of 5th Generation Mobile Communication Technology (5G), industrial internet, artificial intelligence, smart manufacturing, smart construction, and other new technologies or new industries, a new infrastructure has emerged. The scale and the spatial distribution structure of new infrastructure across various regions of China from 2020 to 2022 were measured based on the official annual engineering construction project plans, and the impacts of the digital economy were analyzed by using the spatial Durbin model. The results indicate that the new infrastructure projects are mainly distributed in Eastern, Central, and Southern China. Meanwhile, the Moran’s I index demonstrates a statistically significant positive spatial autocorrelation, indicating that there is a significant trend of spatial agglomeration among these new infrastructure projects. The digital economy is significantly affecting the spatial distribution of the new infrastructure. The impact of the digital economy on the new infrastructure can cross regions and has significant spatial spillover effects, which reflects the virtual characteristic of the new infrastructure and digital economy. Therefore, various regions should coordinate the digital economic development needs of local and surrounding regions to rationally plan new infrastructure projects, ensuring the supporting capacity of infrastructure while preventing over-investment.

To investigate the water film depth (WFD) of super-multi-lane (SML) pavement and its effect on tire-pavement surface interaction, a rainfall-runoff model of SML pavement was developed, and the influence of the number of lanes and pavement slopes was analyzed. Based on the realistic texture of pavement surface, a tire-pavement surface-water film coupling model was established. The contact characteristics between the tire and the pavement surface, as well as the critical hydroplaning speeds under different WFDs, were analyzed. The results show that the WFD of the pavement increases significantly with the increase in lane number. Pavement expanding from three to five lanes in a single direction has the largest increase of the WFD, while the increment decreases as the number of lanes further increases. Increasing the transverse slope from 1.5% to 2.0% results in the most obvious decrease in the WFD. Moreover, increasing the longitudinal pavement slope to 1.0%-2.0% significantly reduces the area of thick water film on the pavement surface. The higher WFD results in a more rapid decrease in the tire-pavement surface contact force as the vehicle speed rises, leading to a lower critical hydroplaning speed. The critical hydroplaning speed of the vehicle decreases by approximately 6 km/h for every 2 mm increase in the WFD.

Invitational bidding is key to encouraging enterprise participation in the technological innovation of public goods. This study develops a contest model that incorporates supplier type (independent or public sector-affiliated), heterogeneous innovation capabilities, and project-specific technical risks. Using limiting-case analysis, we compare two invitational bidding strategies on innovation and efficiency of benefits: inviting only independent suppliers and inviting both independent and affiliated suppliers. We also identify the boundary conditions under which long-term contracts lower the public sector’s obligatory award budget. The key findings are as follows: Competition is strengthened when technically strong independent suppliers bid alongside affiliated suppliers; however, competition weakens when independent suppliers lack technical strength. Long-term supply contracts may not necessarily enhance competition and may diminish suppliers’ motivation for technological innovation. Innovation and benefit improvement are rarely achieved simultaneously. Both goals can be achieved only when independent suppliers have clear technical advantages, making it optimal for the public sector to invite both independent and affiliated suppliers to bid. When innovation projects involve high technical risks or independent suppliers demonstrate strong innovation capabilities, the public sector may reduce the award budget using long-term contracts as incentives to successfully motivate suppliers.

In mobility-as-a-service (MaaS) platforms integrating ridesharing with public transit, each rider-driver pair may have multiple potential matches via different transfer nodes, with users being self-interested with heterogeneous preferences. After generating all feasible integrated matches, a two-sided one-to-one stable matching model is formulated to maximize platform revenue, where each feasible match corresponds to a stability constraint embedding preference information. To solve this model efficiently, an iterative constraint-generation algorithm is designed. It repeatedly solves a restricted master problem to obtain a temporary solution and a subproblem to identify violated stability constraints, iterating until no violations remain. The proposed algorithm can significantly improve computational efficiency. Compared with a centralized matching benchmark with the blocking rate up to 75%, stable matching increases transit usage and user acceptance at the cost of a 31.43% reduction in average platform revenue. Riders experience longer detours with greater cost savings, whereas drivers exhibit the opposite pattern.

Addressing farmers’ financial constraints in the fresh agricultural product supply chain, this study examines a two-stage supply chain involving farmers and e-commerce platforms and explores a hybrid financing strategy that integrates presale financing and multistage crowdfunding. Recognizing that farmers’ presale revenues through e-commerce platforms often fail to cover production costs, this study innovatively constructs a combined financing model of presale and multistage crowdfunding. First, farmers obtain a portion of their start-up capital through presales on e-commerce platforms. Second, crowdfunding is initiated in stages to address the funding gap (planting, growing, and harvesting periods), with the proportion of investor revenue-sharing in each stage dynamically adjusted to balance risk premiums. This study establishes a dynamic planning model to maximize farmers’ profits. The objective function integrates the production volume decision, multistage financing cost, and investor revenue share constraints and solves the model’s equilibrium solution using the Karush-Kuhn-Tucker (KKT) condition and the backward induction method. Numerical simulations show that unlike the traditional single presale model, multistage crowdfunding enhances farmers’ financing capacity while reducing fluctuations in investors’ returns through a dynamic revenue-sharing mechanism. When the number of crowdfunding stages increases to three, the overall efficiency of the supply chain improves relative to single-stage financing, which verifies that stage-by-stage financing enhances fund-matching efficiency. This study provides a theoretical basis for addressing the mismatch problem of “short-term loan and long-term investment” in agricultural supply chain financing, and offers reliable suggestions for e-commerce platforms to design financing portfolio strategies.