Existing methods for audio-visual cross-modal association learning often adopt a dual-stream network structure, but they still face challenges in reducing computational complexity, model light weighting, and efficient feature fusion. To improve model performance and enhance the efficiency of cross-modal learning, a single-stream network-based approach for audio-visual cross-modal learning was proposed. Firstly, preprocessed data from both modalities were fed into a single-stream feature extraction network, where a class-information-based loss function was employed to learn and extract feature vectors from both modalities. Subsequently, attention-based feature fusion was performed on the extracted feature vectors from both modalities. Finally, a combination of cosine similarity algorithm and cross-entropy loss was used to learn the association between the two modalities, thus completing the cross-modal association learning task. Experimental results demonstrate that the proposed method achieves promising performance in audio-visual cross-modal verification, matching, and retrieval tasks, ensuring excellent performance while considering the lightness and flexibility of the network structure.

The stress-strain characteristics of gas hydrate-bearing sediments (GHBS) are essential for achieving safe and secure extraction of marine natural gas hydrate. To quantitatively investigate the influence of hydrate formation on the mechanical properties of GHBS, laboratory tests were conducted under different confining pressures and various hydrate saturations, and it was found that hydrate primarily enhanced the cohesive strength of GHBS, while the internal friction angle remained relatively unchanged. To comprehensively describe the intricate mechanical properties of GHBS, the previously established elastic-plastic constitutive model was employed for predictive purposes. A comparative analysis revealed that the constitutive model exhibited a certain level of applicability. Subsequently, a computer program was developed for the constitutive model by leveraging the UMAT user material subroutine interface provided by ABAQUS. The program was written in FORTRAN language and integrated into ABAQUS, facilitating the implementation of the constitutive model. To validate the effectiveness and stability of the developed model, finite element simulations were conducted on single elements as well as triaxial specimens of GHBS. The results of these simulations confirm the efficacy and reliability of the developed model.

To investigate driving workload on mountainous expressways, a naturalistic driving study was conducted utilizing an eye tracker to capture drivers' eye movement data in a realistic driving environment. Employing the change rate of pupil area, average saccade time, blink frequency, and fixation time ratio as primary indicators, a quantitative driving workload model was formulated through a combined weighting approach. This model aimed to reveal the driving workload evolution mechanism in typical scenarios of mountain expressways such as bridge and tunnel clusters, tunnel clusters, and short distances between tunnels and intersections. A clustering algorithm was applied to determine the classification thresholds for driving workload, thereby identifying high-risk scenarios characterized by heightened workload. The results show that the types of bridges within bridge-tunnel groups, the length of connection sections between tunnel groups, and the proximity of tunnels to interchanges significantly influence driving workload. A positive correlation was observed between driving workload and bridge size, whereas driving workload exhibited a negative correlation with the length of connection sections between tunnel groups and the distance from tunnels to interchanges. The thresholds of high, medium and low intensity levels of driving workload on mountain expressway are 0.54 and 0.26. Scenarios with bridge-tunnel groups composed of large or super-large bridges, tunnel groups with distances less than 300 m, and tunnel-to-interchange sections with distances less than 400 m were classified as high-risk driving workload scenarios. It is advisable to implement an intelligent lighting system within expressway tunnels, establish light-reducing structures at tunnel entrances, and in scenarios where tunnels are located in close proximity to interchanges, consider installing designated lane-changing zones within suitable tunnel sections to facilitate smooth lane transitions.

Addressing the limitation of existing sequence labeling approaches in effectively recognizing nested entities within Chinese electronic health records (EHRs), a novel named entity recognition model that integrates MacBERT and a global pointer network was proposed. Initially, the MacBERT-large pre-trained model transformed the text into context-sensitive dynamic vectors. Subsequently, the fast gradient method (FGM) was employed to generate adversarial samples, which were incorporated into the original vectors and fed into a BiLSTM (bi-directional long short-term memory) network to capture contextual features. To enhance the capture of long-distance semantic features, an attention mechanism was introduced. Finally, a global pointer network model was leveraged to decode simultaneously considering both head and tail feature information, thereby achieving superior prediction performance for medical nested entities. Experimental results demonstrate that compared to the state-of-the-art global pointer model, the proposed model achieves an improvement of 1.8%, 1.37%, and 1.72% in F₁-score on the CCKS2019 dataset and two versions of the CMeEE Chinese EHR dataset, respectively, validating the effectiveness of the proposed approach.

To meet the increasingly refined individual-level traffic management and travel service needs in the new era, a vehicle travel destination prediction method that comprehensively considers temporal and spatial correlation was proposed based on the traditional prediction method based on historical trajectories. Using data from video AI recognition and vehicle satellite positioning, the vehicle stopping points were identified to segment the vehicle's full-day travel trajectories and establish a historical vehicle travel trajectory database. By studying the temporal and spatial characteristics of vehicle travel, a calculation method for the temporal and spatial correlation of vehicle travel trajectories was proposed, and a vehicle travel destination prediction model was constructed using temporal and spatial correlation as weights. Taking the vehicle travel in Futian Central District of Shenzhen as an example, four typical vehicle travel trajectories including private cars and taxis were selected to establish a model prediction accuracy evaluation function. The prediction accuracy of travel destinations for different types of travel and different degrees of trajectory completion was analyzed and compared with the historical trajectory-based prediction method. The results show that the prediction accuracy of travel destinations for different types of vehicles is basically positively correlated with the degree of trajectory completion. When the trajectory completion rate reaches 80%, the accuracy of travel prediction basically reaches over 80%. Compared with the traditional prediction method based on historical trajectories, the prediction method considering temporal and spatial correlation has higher prediction accuracy, especially for taxis services with no fixed commuting travel characteristics. The prediction accuracy of travel destinations has been improved by more than 16%. The research results can better meet the needs of global traffic management.

Due to ignoring the influence of foundation pit unloading effect on the bearing properties of uprooted piles, the accuracy of in-situ test results needs to be studied when the excavation unloading area is large, and the current research on uplift piles mostly reflects the bearing properties of piles based on pile side friction resistance, and there are few studies on the normal pressure of pile-soil interface on pile sides. Compared with the frictional resistance of the pile side, the normal pressure of the pile-soil interface on the pile side can intuitively reflect the influence of foundation pit excavation and unloading on the bearing properties of the uprooted pile. In view of this, the combination of model test and numerical simulation was used to explore the stress response and deformation law of uplift piles under excavation and unloading conditions by changing the variables such as foundation pit excavation range and pile roughness, and then the bearing mechanism of uplift piles was revealed. The results show that the unloading effect of foundation pit excavation is related to the depth and width of the foundation pit, the effective pile length and the pile side roughness, according to the degree of influence: the effective pile length> the depth of the foundation pit> the roughness of the pile side > the width of the foundation pit. When the width of the foundation pit is greater than the critical value, the increase of the width of the foundation pit has little effect on the bearing characteristics of the uplift pile. Under the condition of the same effective pile length, the normal stress of the excavated pile side is proportional to the depth of the foundation pit and inversely proportional to the width of the foundation pit, and the relationship with the pile roughness is small. Under the condition of the same effective pile length, the frictional resistance of the pile after excavation is directly proportional to the depth of the foundation pit and the roughness of the pile, and inversely proportional to the width of the foundation pit.

To study the damage characteristics of rock masses in cold regions under the coupling effect of freeze-thaw cycles and fatigue loads, freeze-thaw cycle tests, nuclear magnetic resonance microscopic tests, and triaxial compression fatigue macroscopic tests under different confining pressures were carried out on single crack marble. The pore structure expansion characteristics of fissured marble after freeze-thaw cycles, as well as the mechanical properties and deformation failure characteristics of fissured marble under freeze-thaw cycles and fatigue loads were analyzed. A single crack marble fatigue damage constitutive model based on Nishihara creep model was established by introducing damage variables under freeze-thaw and fatigue loads. The results showed that with the increase of freeze-thaw cycles, the mass loss rate and longitudinal wave velocity loss rate of fissured marble gradually increased. The nuclear magnetic resonance T₂ curve showed a three peak characteristic, with the whole curve shifting to the right. The second peak gradually connected with the third peak, and the total peak area continued to increase. There was a positive correlation between confining pressure and the fatigue strength of fissured marble. With the increase of confining pressure, the fatigue resistance and ductility of fissured marble increased, and the failure mode of the rock sample gradually changed from local shear failure to overall shear failure. The theoretical curve of the established fatigue damage constitutive model was basically consistent with the experimental data. The research results could provide reference for the prevention and control of rock engineering disasters in cold regions.

In order to meet the requirements of high-voltage DC circuit breakers for rapidity actuation of their actuating mechanisms, a new electromagnetic repulsion mechanism was designed for 40.5 kV vacuum circuit breaker, which had high breaking and closing speeds and was suitable for quick disconnection. It consisted of a coil-plate and a double-coil mechanism connected in series.Firstly, the electromagnetic field simulation was carried out by the finite element method, the feasibility of the mechanism was initially verified by analyzing the electromagnetic repulsion,displacement/time characteristics and velocity/time characteristics. Then, the single-variable method was used to simulate and analyze its motion characteristics. the law of the influence of each parameter on the motion characteristics was obtained, and the optimization parameters were determined. Finally, in order to reduce the tripping bounce of the long-stroke fast electromagnetic repulsion mechanism, an electromagnetic buffer was designed. the effects of buffer current input time and buffer driving circuit parameters on buffer characteristics were analyzed. The results show that the new electromagnetic repulsion mechanism combines the advantages of the fast response of the coil-plate mechanism and the high drive efficiency of the double-coil mechanism,with its short response time, large motion speed and short full stroke time, and meet the need of quick disconnection. Under the parameter conditions of 3 500 μF capacitance and 1 200 V for the driving circuit and buffer circuit capacitance 3 500 μF capacitance and 1 300 V for the buffer circuit, the full travel time of the designedelectromagnetic repulsion mechanism is only 2.18 ms.

Transonic fan rotor blades will suffer from morphological decay problems such as leading edge erosion in actual operation, and the flow field structure in the tip zone will change and then induce aerodynamic performance degradation. The effects of leading edge erosion on the tip leakage flow of the fan rotor were investigated from constant numerical computation in this paper. The results show that at the stall point of the eroded blade, its isentropic efficiency, total pressure ratio and mass flow rate decrease by 4.3%, 0.43% and 5.63%, respectively, and the leading-edge erosion also causes a decrease in the stabilized operating margin by 0.69%. For the flow in the tip zone, the leading-edge erosion causes an increase in the entropy increase area and intensity in the flow surface of S1 at the tip zone and the flow surface of S3 in the exit region, and the formation of a low Mach Number flow region in different blade heights; the leakage flow structure under the leading edge erosion undergoes malignant changes, the demarcation point of the main leakage flow and the secondary leakage flow is advanced, carrying more fluid for secondary leakage, and the leakage vortex deflection angle of the eroded blade reaches 30° from 15°, hitting the leading edge of the blade directly and causing flow blockage.

The tension at the suspension point of transmission towers is an important parameter for line renovation and collapse accident warning. Although airborne LiDAR can provide high-precision spatial information for transmission line scenes-LiDAR point cloud data, lacks timely feedback on the operational status information of elements (such as tension of suspension points). A tension extraction method for tower suspension points based on LiDAR point clouds was proposed, and a new approach for studying micro physical parameters of transmission line spatial information was designed. Firstly, a three-dimensional spatial model of the transmission line was reconstructed from point cloud. Then, a mapping model between the spatial curve model and the horizontal stress of the transmission line was constructed. Finally, based on the tension vector relationship of adjacent transmission lines, the tension at the suspension point of the tower was extracted. This method reconstruct the three-dimensional model of transmission lines and extract the tension of tower suspension points in real-time operation based on LiDAR point cloud. Compared with finite element analysis method, the maximum relative error is 3.26%. The research can be extended and applied to the safety inspection and state parameter detection of large-scale transmission lines.