Traditional machine vision detection methods suffer from low accuracy in identifying small-scale defects. To address this, a nondestructive identification method for steel surface defects is proposed based on an enhanced version of the fifth version of the You Only Look Once(YOLOv5)algorithm. In this improved approach, the Res2Block module is incorporated into the backbone of the YOLOv5 algorithm to expand the receptive field and improve computational efficiency. Additionally, the recursive gated convolution structure is fused into the neck of the YOLOv5 algorithm to further enhance the computational performance of the surface defect identification method. To validate the effectiveness of the proposed method, a series of ablation experiments were conducted using different module combinations. These results were then compared with those obtained through other object detection methods. This comparison reveals that the proposed method achieves a mean average precision of 67.8% and an F₁-score of 86.0% in steel surface defect identification. When compared with the original YOLOv5 algorithm, the proposed method exhibits superior performance, particularly in the identification of small-scale steel surface defects. Furthermore, it also surpasses other object detection methods, such as SSD, YOLOv3, YOLOv5-Lite, and YOLOv8, demonstrating significant improvements in computational accuracy.

A novel type of tuned viscous mass damper(TVMD)device incorporating eddy current damping and metal springs as its damping and spring elements, respectively, was developed. First, dynamic tests of the TVMDs were conducted to investigate their dynamic characteristics. Subsequently, four methods were proposed to identify the TVMD parameters from the test data: the peak point fitting method, hysteretic curve fitting method, time-history fitting method, and transfer function fitting method. The dynamic test results indicate that the TVMD exhibits the inertial mass amplification and damping enhancement effects. The spring and inerter elements demonstrate ideal linear behavior, while the damping element exhibits nonlinearity, primarily owing to the nature of eddy current damping and inherent friction in the TVMD device. The parameter identification results indicate that all four methods can reasonably determine the TVMD parameters. The transfer function fitting method can provide an equivalent damping coefficient useful for tuning design, while the other three methods can identify parameters of nonlinear damping models. The hysteretic curve fitting and time-history fitting methods exhibit improved accuracy in parameter identification, while the peak point fitting and transfer function fitting methods exhibit higher computational efficiency.

To address the issue that conventional disturbance observer(DO)design did not consider actuator saturation, which is prevalent in practical systems, a DO-based control(DOBC)strategy with anti-windup compensation is proposed. First, the reason of windup phenomenon in the conventional DOBC when the actuator is saturated is studied. Then, an anti-windup compensator is designed by minimizing the performance index, and patched to the DO so that the modified DOBC can effectively handle actuator saturation. Finally, local asymptotic stability analysis is performed on the resulting closed-loop systems. Comparative simulation results show that when there is actuator saturation, the proposed method has smaller errors in position tracking and disturbance estimation, and the designed compensator can maintain the DO states to be as close as possible to those without actuator saturation. This verifies that the proposed method is superior in anti-disturbance and anti-windup.

To enable accurate vessel recognition for bridge collision avoidance and early warning, an image dataset for vessels in bridge channels is established using cameras and data augmentation. This dataset includes complex scenarios such as long distances, multiple targets, and low visibility. Subsequently, the you-only-look-once version 5(YOLOv5)model is employed as the basic detector, and several modifications are applied to its network structure. Key enhancements involve replacing C3 modules in the backbone network with C2f modules, integrating the squeeze-excitation attention mechanism into the feature fusion network, and optimizing the prior anchors of the dataset using the K-means++ clustering algorithm. Finally, the modified model undergoes training and validation using PyTorch as the deep learning framework. Results demonstrate that the mean average precision for crucial vessels in the modified YOLOv5 model reaches 99.4%, representing an 11.1% improvement compared to the original YOLOv5 model. Additionally, the inference speed is measured at 102 frame/s. The established YOLOv5 model is a reliable and efficient cornerstone for warning against vessel-bridge collisions in complex navigable scenes.

To explore the benefits and potential of electricity and hydrogen as alternative fuels for regular buses, a mixed-integer planning model was constructed to determine the schedule optimization scheme for bus fleet replacement. The model was based on the comprehensive analysis of carbon emissions and the total cost of ownership from a life cycle perspective. Using actual operational data of buses powered by diesel, natural gas, hybrid, plug-in electric, and hydrogen fuel cells, the effects of uncertainty in the power mix, acquisition cost, hydrogen production, and hydrogen usage cost on the fleet replacement schedule were explored. The results reveal that plug-in electric buses are currently the optimal choice for bus fleet replacement. Given the current level of vehicle technology and hydrogen production, hydrogen fuel cell buses(HFCEBs)are advisable during bus fleet replacement. Until the production of blue or green hydrogen becomes commercially viable, promoting HFCEBs on a large scale by extending financial subsidies is not recommended. The proposed method can help authorities identify optimized bus fleet replacement options under specific constraints and desired objectives to promote green and sustainable development.

To investigate the performance of Lamb wave monitoring of fatigue cracks at the floor beam cutout of orthotropic steel bridge decks, two full-scale specimens were produced and subjected to cyclic loading. The cutout cracks initiated during the tests were continuously monitored by Lamb wave sensors. Crack growth was estimated based on standardized wave features extracted from the received waves. Finite element models with typical regions were also established and validated by the experimental results. Parametric studies were conducted to determine the optimal sensor arrangements for monitoring cutout cracks, considering various parameters, such as crack lengths and sensor locations. The experimental results indicate that the standardized wave features increase when cracks and wave propagation paths are perpendicular, whereas the standardized wave features decrease when cracks and wave propagation paths are parallel. The parametric results reveal the optimal sensor arrangements for crack monitoring in the floor beam cutout regions, i.e., prearranging one excitation sensor and four reception sensors within the span of a typical floor beam cutout region. The excitation sensor should be placed at a distance of 50 mm from the cutout, whereas the reception sensors should be arranged at a distance of 50 mm from the cutout or 50 to 100 mm from the deck plate.

To accurately predict the settlement of soft soil foundations under graded surcharge preloading, the Asaoka method was improved. Considering the change in the soil consolidation coefficient with load, the slope of the settlement prediction line was modified, and a prediction calculation method for the prediction line intercept was used. The improved Asaoka method considered the influence of consolidation stress on the consolidation coefficients and nonlinearity of soil consolidation. The reliability of the improved method was then verified through indoor consolidation tests and field observation. The results demonstrate that the consolidation coefficient C_v is not a constant, and it satisfies the relationship with the consolidation pressure P: C_v=aln(blnP), which has an error of less than 5%. The settlement prediction lines at various stress levels are not parallel, and the improved method exhibites a lower error rate than the original Asoaka method. The improved Asaoka method offers higher prediction accuracy than the traditional method and can reliably predict the settlement of soft soil foundations during graded surcharge preloading.

To explore the design and safety performance of super high-rise connected structures under the combined action of multiple disasters, taking Suzhou Supertall as an example, a vulnerability analysis is conducted under the combined earthquake-wind actions. First, the structure’s finite element model is established. Then, vulnerability assessments are conducted under individual earthquake and combined earthquake-wind actions. Finally, the response law of the structure is obtained. Results indicate that when exposed to combined earthquake-wind actions, the structure’s vulnerability increases with the earthquake and wind intensities, and the seismic action dominates structural damage. The probabilities of moderate, severe, and collapse damages are higher under the combined earthquake-wind actions than those under individual earthquakes. When the wind speed reaches 40 m/s, the probabilities of the structure reaching three failure states under rare earthquakes are 99.77%, 91.56%, and 46.54%, respectively, representing an increase of 1.11%, 10.73% and 14.65% compared with those under rare earthquakes alone and an increase of 0.27%, 6.26% and 14.34% compared with those of a typical high-rise connected structure under the same combined action of disasters.

Through comprehensive data collection, along with the coarse aggregate mechanical index, fractal dimension, and British pendulum number(BPN), a pavement friction prediction model was proposed on the basis of backpropagation neural networks(BPNNs)and support vector machine(SVM). An accelerated attenuation test was conducted to examine the antiskid performance of the asphalt mixture and aggregates at different wearing cycles. Subsequently, BPN was fitted using an exponential model. Gray relational and correlation analyses were performed to evaluate the factors influencing pavement skid resistance. According to the principal component analysis results, six schemes were prepared for the training, validation, and testing of BPNN and SVM algorithms. Test results indicate that different aggregates exhibit different antiskid properties. Quartz sandstone is the most suitable, followed by basalt and limestone. The polished stone value has the highest correlation with the attenuation model of asphalt antiskid performance. BPNN is more stable, with an R² value of approximately 0.8.

To develop comprehensive similarity indicators for assessing the fitting degree of hysteretic curves shape descriptors, widely used in image feature extraction, are introduced. A specific process is proposed to delineate the formation of indicators based on these descriptors, enabling the calculation of curve similarity between numerical simulation and experimental data. Following this process, an indicator is devised based on shape context. First, the similarities of the hysteretic loops in the numerical simulation curve are calculated. Subsequently, the weighted sum of these similarities is calculated to derive the similarity of the entire curve. To verify the effectiveness of the indicator, the Bouc-Wen model is utilized to conduct a numerical simulation study. Five parameters of the model are adjusted, resulting in the formation of 51 numerical simulation curves. Similarities of the curve, along with errors in the force peak point, energy dissipation, and stiffness, are calculated. The results show that the absolute values of Spearman’s correlation coefficients between similarity and errors all exceed 0.78, which has a strong correlation, thus verifying the feasibility of the indicator and the effectiveness of the process for forming the indicators.

To assess the combined risks of long-span suspension bridges under continuous wind loads and occasional earthquakes, a risk assessment framework for cross-sea suspension bridges based on improved Bayesian networks was proposed by combining the quantitative analysis of the structural damage probability and the qualitative assessment of the damage consequences during bridge operation. First, the damage degree of each component was obtained according to the characteristics of the suspension bridge and the results of wind and earthquake analyses. Then, the failure probability of the bridge structure was calculated using the theory of structural reliability. Finally, the risk assessment model of the suspension bridge based on improved Bayesian networks was proposed to evaluate the risk during bridge operation. The results show that considering the varying impacts of different bridge components, the bridge damage level can be categorized into four degrees based on its disaster resilience. Taking the Lingdingyang Bridge as an example, the maximum risk level under multihazard risks is level 3 according to the proposed method, which requires traffic restrictions and maintenance. Therefore, this method can guide the emergency management strategy of sea-crossing bridges in response to multihazard risks.

To solve the problem that existing methods have difficulty in accurately obtaining the spatiotemporal distribution of vehicle loads on bridges in complicated traffic scenes, a spatiotemporal location identification method for vehicle loads based on multi-view information fusion is proposed. First, the vadYOLO-StrongSORT model is developed to detect and track vehicles simultaneously in a single view. Furthermore, based on image calibration and cross-view vehicle matching, an adaptive weighted least squares method is used for multi-view information fusion to correct the vehicle trajectory. Finally, the spatiotemporal distribution of axle loads is reconstructed by combining vehicle trajectories with axle configurations. The performance of the proposed method under typical traffic conditions is evaluated using model tests. The results show that the multi-view information fusion method significantly improves tracking stability, localization accuracy, and anti-occlusion performance compared with the single view-based vehicle location identification method. In the lane-changing scenes, the highest average localization error of the proposed method is less than 2.0 cm, which is significantly better than the 17.0 cm of the single-view method. In multivehicle occlusion scenes, the proposed method achieves a vehicle capture rate of up to 100%, compared with a maximum of only 72.5% for the single-view method. Meanwhile, vadYOLO-StrongSORT achieves the highest identification accuracy in the experiment compared with other detection and tracking models.

The limitations of conventional cable force optimization methods, which fail to automatically optimize and consider the overall performance of the bridge structure, as well as the drawbacks of extensive calculations, lengthy processing time, low efficiency, and slow convergence speed, when combined with intelligent optimization algorithms, should be addressed. Ansys and Matlab are used as the structural calculator and master control programs, respectively, with the minimum bending moment energy as the control objective.Moreover, the influence matrix and elite retention strategy are incorporated into the genetic algorithm to optimize the cable force during the bridge formation stage. This method can simultaneously account for the force characteristics of the main girder and pylon. Utilizing the influence matrix, the issue that each generation requires finite element evaluation can be resolved, thereby drastically reducing the amount of calculation. In addition to capitalizing on the benefits of the conventional influence matrix method, the proposed approach considers the iterative process of parameter selection and permits the addition of special constraint requirements to critical sections of the structure, thereby enhancing the realism of the optimization procedure. Furthermore, the introduction of the elite retention strategy enhances the convergence speed and stability of evolutionary iterations. Finally, a practical engineering application is utilized to validate the viability of the proposed method.

A photoexcited switchable single-band/dual-band terahertz metamaterial absorber with polarization-insensitive and wide-angle absorption is reported. The function switching is realized by modulating the conductivity of the photosensitive GaAs embedded in the resonator, and the surface currents at different GaAs conductivities are extracted to physically explain the absorption mechanism of the metamaterial absorber. The results show that the absorber can realize switching from dual-band absorption at 0.568 and 1.442 THz with 99.08% and 99.56% absorptivity, respectively, to a shift single-band absorption at 0.731 THz with 95.43% absorptivity. The device has an intensity modulation depth of 61.4% and a frequency tuning bandwidth of 60.6%. With these values, the device can be used to fabricate intensity modulators and frequency-selective absorbers in the terahertz band. In addition, the proposed absorber exhibits polarization-independent and wide-angle absorption for transverse electric(TE)and transverse magnetic(TM)polarization waves. The realization of tunable metamaterial absorbers offers opportunities for mature semiconductor technologies and potential applications in active terahertz modulators and switchers.

To solve the problem of heavy artifacts in ultrasonic images, a novel ultrasonic imaging method is presented using a sum of sinc(SoS)kernel for eliminating the artifacts caused by the diffusion of isacoustic path, signal tail, or noise simultaneously. First, the envelope of ultrasonic echo is obtained and passed through a SoS kernel, then the signal is sampled at equal intervals determined by the echo signal information degree, and the Fourier transform is applied to the discrete sampling data to obtain the Fourier coefficient sequence. After that, the spectral estimation algorithm is used to estimate the parameters of the ultrasonic echo signal and reconstruct the echo signal using the estimated parameters. Finally, the ultrasonic image is obtained by calculating the acoustic field using the reconstructed echoes. Experimental results show that the image artifacts are effectively removed using focused and straight probes to test straight slot defects and through-hole defects, respectively. Compared with the B-scan images, the peak signal-to-noise ratios reach 24.306, 23.213, 15.074, and 16.444 dB, and the structural similarity indexs are 0.931, 0.932, 0.746, and 0.773, which indicates that the quality of the defect images is greatly improved using the proposed method.

To accurately identify sensor faults caused by complex environmental conditions and ensure that structural health monitoring systems correctly perceive the structural state, a self-detection method for sensor nodes based on mean shift and sliding window techniques was proposed. The self-detection method comprises two stages, i.e., fault prescreening and fault self-detection. During the fault prescreening stage, the method rapidly identifies potentially abnormal data using the quartile method combined with the sliding window technique, significantly improving the efficiency of the method. During the fault self-detection stage, the method employs the mean shift algorithm to perform adaptive clustering of the abnormal data, effectively detecting various faults. Data from the Canton Tower were used to test the effectiveness of the method by setting four types of sensor faults, i.e., offset, drift, gain, and stuck. Then, the proposed method was compared with extremely randomized trees, random forests, support vector data description, and one-class support vector machines. Results show that the proposed method can detect the four aforementioned faults with high accuracy and computational efficiency.

To facilitate the contact analysis of involute cylindrical gears, a meshing contact analysis and tooth profile modification algorithm program for involute cylindrical gears is designed. The involute spur gear parameter equation is employed to accurately model the radius of curvature of the tooth profile contact point, and the Hertz contact formula is enhanced to solve the gear meshing contact stress considering the actual load. To reduce the impact of gear meshing, five tooth profile modification methods and their contact stress analysis methods after modification are given in the algorithm program. Compared with the calculations by MASTA, the industry-recognized gear design simulation software, the algorithm program has a higher accuracy in solving the contact stress and the contact stress distribution after tooth profile modification accords with the MASTA calculation results, and the average error is about 5.04%, which confirms the rationality of the algorithm program.

To investigate and design an efficient blind detection method for third-party scenarios, a third-party efficient physical downlink control channel(PDCCH)blind detection method was proposed based on polar decoding metric selection. This method comprised two main components: the study of the polar decoding algorithm, which introduced a polar decoding metric based on downlink control information(DCI)length and proposed an improved third-party blind detection method based on polar decoding metric selection; and the investigation of the PDCCH blind detection algorithm, which introduced a reordering blind detection algorithm. The enhanced polar decoding algorithm and reordering blind detection algorithm were organically combined to present an efficient PDCCH blind detection method for third-party scenarios. The proposed method was validated and analyzed using a 5G PDCCH blind detection simulation link on the MATLAB platform. The results show that the proposed method effectively reduces the number of PDCCH blind detections and the count of DCI candidates while enhancing blind detection efficiency and ensuring target capture accuracy.

To tackle the issue of high cost and large volume for offshore wind power generators, a novel dual-rotor dual-stator permanent magnet synchronous generator(DRDS-PMSG)is proposed. The equivalent magnetic circuit model of the generator is established, finite element analysis is performed to evaluate the electromagnetic characteristics and coupling effect, and some simulation results are verified through experiment. The simulation analysis results show that three typical equivalent magnetic circuits exist with changed relative positions between the inner and outer magnets, and the equivalent reluctance of the coupling region can be described using a coupling coefficient. The coupling effect of inner and outer machines is revealed by electromagnetic characteristics, including cogging torque and electromagnetic torque. The peak-to-peak values of the cogging torque of inner and outer machines are 0.52 and 1.64 kN·m, the average values of electromagnetic torque are 11.65 and 27.09 kN·m, and the torque ripples are 6.02% and 4.12%, respectively. In general, a coupling effect exists between the inner and outer machines; however, the coupling effect is effectively reduced by the flux barrier.

To address the challenge of solving free vibration problems in beams with uniform cross-sections, beams with variable cross-sections, and Euler-Bernoulli beams with concentrated masses, an innovative method combining the Rayleigh method and the Monte Carlo method is introduced. This dual-method strategy offers a novel solution by first discretizing the continuous beam structure model, followed by employing the Monte Carlo method to determine the vibration modes of the beam structure. Subsequently, these identified vibration modes are integrated into the Rayleigh method to calculate the fundamental frequency and vibration modes. The process involves a meticulous comparison with the minimum value obtained during calculations to ensure the satisfaction of the convergence condition. The results show that this combined method achieves a maximum error of 10% or less in predicting the fundamental frequency across different calculation models. This accuracy level is well within acceptable engineering requirements. The control parameters for accuracy and time can be easily adjusted to meet various needs. The method, which is simple in theory and widely applicable, enables the quick and precise determination of fundamental frequencies and vibration modes for diverse beam structures.

A novel composite plate made of carbon fiber-reinforced polymer(CFRP)grids and bamboo scrimber, termed CBCP, was developed to enhance the mechanical properties of bamboo scrimber and reduce its anisotropic defects. To investigate the mechanical behavior of CBCP and assess the influence of its composite fabrication method, uniaxial tensile tests were performed on eight sets of CBCP tensile specimens, and pull-out tests were conducted on eighteen groups of CBCP pull-out specimens. Three similar constitutive models were selected to depict the bond stress-slip curves of the pull-out specimens. The critical anchorage lengths between CFRP grids and bamboo scrimber were determined. The effects of the number of CFRP layers, characteristics of transverse CFRP bundles, and anchorage length on the bonding performance between CFRP grids and bamboo scrimber were analyzed. The results reveal that integrating a CFRP grid can increase the tensile strength of bamboo scrimber, both parallel and perpendicular to the grain, by 35.77% and 135.20%, respectively. This enhancement effectively reduces the anisotropic defects of bamboo scrimber. The increase of the number of grid layers and the grid spacing can enhance the tensile performance of CBCP. With the increase of the anchorage length, the average interface bonding strength decreases exponentially while the peak load slip decreases linearly.

Three pricing strategy models—free, charge, and cash subsidy—are constructed for content platforms in a multilateral market based on the game theory. The optimal pricing strategy for a platform is identified by comparing the parameters under each pricing strategy. The results reveal that ad interference cost and ad marginal revenue affect a platform’s pricing strategy selection and the cash subsidy amount. The cash subsidy strategy is used when both are within a certain range of thresholds; the charge strategy is adopted when the ad interference cost is very high; and the free strategy is adopted in other cases. In addition, under the cash subsidy strategy, the amount of cash subsidy is negatively correlated to ad interference cost and positively related to ad marginal revenue. Under the same conditions, adopting the cash subsidy strategy is better for all stakeholders and social welfare than the other two pricing schemes. Moreover, ad marginal revenue affects some parameters in the cash subsidy strategy and the free strategy in opposite directions.

Asphalt mixtures were prepared in the laboratory using reclaimed asphalt pavement(RAP)at a 40% mass concentration to investigate the homogeneity characteristics of RAP during the mixing process, taking into account the mixing duration and aging rate. Specifically, titanium dioxide powder was incorporated into RAP binders as an identification tracer. Homogeneity indices were subsequently established to evaluate the homogeneity characteristics of asphalt mixtures. Computed tomography(CT)was used to create digital images of asphalt mixtures, and Mimics software was used to identify the components. Finally, a homogeneity evaluation system was established to assess the uniformity of both the components and RAP agglomerates. The results indicate that homogeneity is mostly influenced by the distribution of coarse aggregates, followed by asphalt binders and air voids. Homogeneity improvement is hindered by the formation of new agglomerates during the mixing process, whereas it is increased by prolonging the mixing time. A homogeneity evaluation system based on components and agglomerates can effectively reveal the principles governing homogeneity characteristics and provide a reference for the construction of high-quality recycled asphalt pavement.

A comprehensive model based on continuum theory is adopted to conduct the parametric analysis of the primary resonance of the nonlinear vibration of spatial cable suspension bridges. This model can simultaneously account for the geometric nonlinearity of both the vertical motion of the deck and the vertical-horizontal motion of the cable. Based on this model and the multiple scale method(MSM), the modulation equations of the primary resonance responses are derived for spatial cable suspension bridges. Nonlinear coefficients in the modulation equations are determined to have notable influences on the maximum response amplitude of the primary resonance of the system and the hardening or softening characteristics of the investigated vibration mode. Meanwhile, system parameters, such as the inclination angles of the main cable and hanger, the sag-to-span ratio of the cable, and the tensile stiffness ratio between the deck and cable, can notably influence the nonlinear coefficient. The dynamic properties of the system can change dramatically in the form of sudden changes in the nonlinear coefficient of the symmetric vibration of the deck and cable if the parameter is located near the singularity, which should be avoided in the design of the system. This study can provide reference for the design of the bridge structure.