Particle profile control and blockage is recognized as an important method for enhancing oil recovery. The migration and deposition characteristics of particles in porous media are understood to facilitate the optimization of particle preparation, thereby improving compatibility with reservoir pore throats and blocking efficiency. Factors such as particle concentration, particle size, porous medium structure, particle size ratio, and fluid parameters within the medium were reviewed for their effects on migration and deposition. Research results from various simulation methods, including simplified geometry, mesoscopic simulation, lattice Boltzmann method-discrete element method (LB-DEM) and computational fluid dynamics-discrete element method(CFD-DEM) were summarized. It is indicated that the critical value of the particle size ratio influences the deposition location and blockage degree in porous media. Different particle sizes are subjected to significant differences in forces, with larger particles being notably affected by hydrodynamics, gravity, and fluid flow rates. The fluid flow model within porous media is not yet fully unified. However, the Brinkman-Forchheimer-Darcy model is noted for its strong applicability. The CFD-DEM method, approached from a microscopic perspective, has validated the flow-solid coupling of migration and deposition within the medium, providing a basis for profile control schemes in heterogeneous reservoirs.

Under the guidance of the “14th Five-Year Plan” and the “Dual Carbon” goals, construction materials face significant challenges, particularly as the adaptability and accuracy of traditional concrete performance prediction models are questioned. Recently, machine learning (ML) has demonstrated high accuracy and efficiency in predicting concrete performance. The research progress of ML in this field was systematically reviewed, focusing on its applications in mechanical properties, mix design, and durability, while identifying its limitations and proposing improvement strategies. CiteSpace software was used to analyze the current state of ML research in construction engineering, examining publication volume, research hotspots, and trends. This analysis offers valuable reference for future researchers, aiding in the effective application of ML technology to drive innovation in construction materials and support environmental sustainability goals.

In rock image recognition, achieving rapid and accurate identification of rocks is crucial for the digitalization of rocks. Among the challenges faced in intelligent rock recognition is the issue of image blurring caused by environmental factors such as lighting and humidity. In light of this, a novel deep learning approach (MobileNetV3-small-RegNetX) was proposed for rock image recognition, which is suitable for scenarios with limited resources such as mobile devices. Building upon the RegNet network, transfer learning methods, combining the advantages of the MobileNetV3 residual structure with squeeze-and-excitation (SE)modules was employed to effectively optimize feature extraction and network structure, leading to a significant improvement in detection speed. To validate the accuracy of this approach, comparative experiments were conducted between the new model and current mainstream lightweight models (DenseNet and ShuffleNet). The results demonstrate that the new model proposed exhibits high precision (82.15%) and fast processing (0.06 GFLOPs). Additionally, the model demonstrates good adaptability to environmental factors such as lighting and humidity-induced image blurring.

In order to reveal the characteristics and geological significance of the main micronutrient geochemistry in the Carboniferous-Permian Taiyuan Formation in Ningdong coalfield which deposited in marine-land transition coal-accumulating environment, the X-ray fluorescence(XRF) spectrometry, inductively coupled plasma-mass spectrometry(ICP-MS), macerals identification and industrial analysis were used to study the source, occurrence and environmental significance of trace elements in coals. The results show that the content of major and trace elements in Carboniferous coals of Ningdong coalfield are varies greatly, except the Fe₂O₃ content is lower than the mean content of China coals. The Rb(5.1) is enriched in Weierkuang while other elements are slightly enriched in different micronutrient. The major and trace elements in coal are mainly occurs in clay minerals, and come from the supply of terrigenous clasts and the combination with organic matter and authigenic minerals in water-soluble state and ion state. The ratio of TiO₂/Al₂O₃indicate that the minerals in the coal are mainly derived from felsic clastic rocks, and the positive correlation between CaO and MgO indicates that they coexist in the form of dolomite. The ratio of Sr/Cu and CaO/(MgO·Al₂O₃) and the other parameters indicate the coal-accumulating period is in a warm and humid environment, it is inferred that the temperature is between 15 ℃ and 30 ℃, the ratio of Sr/Ba indicates that the coal-forming marsh is a brackish water environment, which indicates that it is affected by seawater. The parameters of Cu/Zn, V/(V + Ni), Ni/Co and V/Cr revealed that the peat-formed bog was in an anaerobic redox environment.

Dafuling(DFL) deposit is a part of the Mingyuefeng ore field in eastern Hunan Province, China, is a typical perigranitic uranium deposit discovered recently. The characteristics of the ore minerals and the features of rare earth elements in this deposit have not been previously documented. In order to further elucidate the characteristics of ore minerals and rare earth elements (REE), as well as explore their indicative significance for uranium metallogenesis. Herein, uraninite, the primary ore mineral in the deposit, was investigated via scanning electron microscopy and electron probe microanalysis. Additionally, laser ablation-inductively coupled plasma-mass spectrometry(LA-ICP-MS) was used for the first time to determine the in situ the REE characteristics of uraninite. Uraninite exhibits a distinct fractionation between light rare earth elements(LREE) and heavy rare earth elements(HREE), while displaying a negative Eu anomaly. The (La/Yb)_N ratio exceeds 1, indicating a significant enrichment of LREE. Both major elements and REE in uraninite suggest its formation within a hydrothermal environment at temperatures ranging from moderate to low, below 350 °C. Consequently, DFL deposit can be classified as a typical hydrothermal vein-type uranium deposit. The REE serve as indicators of the transition of ore-bearing hydrothermal fluids from high salinity to low salinity, and the hydrothermal environment gradually shifts from a reducing state to a weakly oxidizing state. These observations suggest that the ore-bearing hydrothermal fluid responsible for metallogenesis originating from the deep crust or lithospheric mantle, ascended along regional deep faults, and subsequently underwent a series of physical and chemical transformations, eventually accumulated mineralization in suitable locations within DFL deposit.

The Permian Wutonggou Formation in Dixi area of Junggar Basin has huge potential for oil and gas exploration. Based on the latest seismic data, combined with thin section data, drilling and logging data, rock physical properties and physical parameters, using seismic forward modeling, wave impedance attributes and waveform clustering attributes, the basic characteristics of the reservoir were characterized, the seismic waveform identification method of the upper and lower sand groups in the first member of Wutonggou Formation was clarified, the thickness distribution law of the upper and lower sand groups was described, and the sedimentary facies development characteristics of the upper and lower sand groups were clarified. The results display that the sandstone of Wutonggou Formation is mainly lithic sandstone and feldspar lithic sandstone, which belongs to low porosity and low permeability reservoir. The seismic waveform characteristics of sand bodies are significantly affected by seismic resolution, sand body thickness, mudstone interlayer thickness, sand body superposition relationship and underlying lithology. The sand body distribution regular pattern based on the interpretation of the sand body waveform characteristics of the forward model is highly consistent with the average wave impedance attribute distribution regular pattern. The study area develops delta front underwater distributary channel microfacies, estuary dam microfacies, underwater tributary bay microfacies and sheet sand microfacies. There are great differences in the development characteristics of sedimentary facies between the upper and lower sand groups.

Landslide disasters pose a serious threat to residents’ lives and socio-economic development. Taking Xinyuan County as the study area, 17 landslide influencing factors were selected as the initial factor set. Through multiple collinearity analysis, 10 landslide factors were screened and an evaluation index system for landslide susceptibility in the study area was constructed. The landslide susceptibility was then evaluated based on three typical models: logistic regression (LR), support vector machine (SVM), and random forest (RF). The evaluation results of the model were compared and validated using the area under curve (AUC), landslide ratio, and field investigation under the receiver operating characteristics(ROC) curve. The results show that low-susceptibility areas are mainly concentrated in the valley plain of the Kongnais River, where the terrain is flat and the landslide susceptibility is relatively low. High-susceptibility areas are mainly located in the northern part of the Kongnais River valley, the Awulale hilly area, and the watersheds on both sides of the southern Yishikelike Mountains and Nalati Mountains, as well as the area south of the Qiafu River, where the terrain is complex and varied, leading to a higher susceptibility to landslides. Among the three evaluation models, the SVM model performs the best, with an AUC value of up to 0.985, indicating its high accuracy in landslide susceptibility assessment. Furthermore, the high-susceptibility areas identified by the SVM model have a high density of landslide points, accounting for 86% of the total, further validating its effectiveness in landslide susceptibility assessment. Based on the above results, the SVM model is more reasonable than the other two alrorithms in assessing landslide susceptibility in Xinyuan County, providing a scientific theoretical basis and reference for landslide prevention and control in the region.

Landslide geological hazard susceptibility assessment is an important means of hazard prevention and reduction. The selection and optimization of susceptibility assessment model is very important. Sinan County was selected as the study area, and 16 assessment factors such as elevation, slope, curvature, lithology, land use, and average annual precipitation were selected. Frequency ratio (FR) model was coupled with support vector machine (SVM) model and random forest (RF) model. Grid search method was introduced to obtain the optimal parameter combination of SVM model, RF model and their coupling model for model training. Finally, SVM, RF, FR-SVM and FR-RF models were constructed to predict landslide susceptibility in the whole study area, and receiver operating characteristics (ROC) curve was performed verification. The results show that compared with the single machine learning model, the coupled machine learning model has more landslide hazard samples fall in the high zone and the very high zone, and has higher accuracy. In the single model, more landslide hazard samples in the RF model fall in the high zone and the extremely high zone. In the coupled model, more landslide hazard samples in the FR-RF model fall in the high zone and the very high zone, and no hazard samples points in the FR model and the FR-RF model fall in the very low zone, indicating that no matter the single model or the coupled model, The performance of RF model is better than that of SVM model. The AUC values of ROC prediction curves of the four models are 0.831 6, 0.843 9, 0.864 4 and 0.910 4, indicating that the coupling model combined with FR model and RF model has a higher accuracy, and this model is more suitable for the assessment of landslide susceptibility in Sinan County. The assessment results can provide some reference for hazard prevention and reduction of local landslide geological hazards.

In order to explore the effects of different internal fixation systems on the biomechanical characteristics of the spine after orthopedic idiopathic scoliosis, a theoretical basis for the improvement of the internal fixation system was provided from the perspective of biomechanics. Based on reverse engineering, topology optimization and finite element modeling techniques, the finite element model of idiopathic scoliosis was established by taking actual cases as examples. The personalized fusion device was designed. Two kinds of internal fixation systems were established, namely full fixation and interval fixation. To simulate idiopathic scoliosis surgery and compare the biomechanical differences between spine and internal fixation system under different physiological conditions. The results show that the average stress of cortical bone and cancellous bone is increased by 17.19% and 12.37%, respectively, compared with that of interlocking nails. The maximum equivalent stress of fibrous annulus matrix and nucleus pulposus is increased by 1.78% and 1.1%, respectively, compared with that of full nailing. The maximum equivalent stress of pedicle screws is 11.64% higher than that of interlocking screws. The average stress of interbody fusion is increased by 6.15% compared with that of interbody fusion. In conclusion, compared with the interspaced nailing scheme, the total nailing scheme is better in vertebrae safety, but the risk of screw slip and screw loss is higher. Compared with the total nailing scheme, the interstice nailing scheme has better spinal fusion effect and effectively alleviates stress occlusion, but the incidence of bone hyperplasia is increased.

With the rapid development of infrastructure projects in China, the use of anchor bolts in mining, geology, and tunnel engineering continues to increase. The non-destructive testing of anchor bolt quality is crucial for enhancing the stability and safety of engineering projects. Based on the stress wave detection method, wavelet threshold functions and STA/LTA algorithms were employed to comprehensively evaluate anchor bolt engineering. A software for non-destructive testing and intelligent analysis of anchor bolt anchorage was developed, integrating signal filtering and acquisition of anchorage parameters. Through numerical simulation analysis of anchor bolt anchorage, the lengths of anchor bolts were calculated using both manual picking and software-based arrival time extraction. The results show that the software-calculated anchor bolt lengths have an overall error controlled within 5%, offering higher precision than manual picking, which is significant for improving the safety and stability of project engineering.

In recent years, multiple exploration breakthroughs have been made in the Maokou Formation of the Sichuan Basin. Currently, the overall exploration level is low.The type and distribution of sedimentary facies are still unclear. The characteristics and formation mechanism of the reservoir are unclear. Techniques such as core observation, pore permeability testing, conventional and cast thin section identification were used to clarify the sedimentary and reservoir characteristics of the 2nd Member of the Maokou Formation in the front of the Longmen Mountains in western Sichuan and to explore the main controlling factors of high-quality reservoirs.Research shows that the 2nd Member of the Maokou Formation in the front of Longmen Mountain in western Sichuan develops in a northwest southeast direction. The southern and middle sections of Longmen Mountain are the development areas of high-energy shoals of the lower sub section of the second section of Maokou Formation. The middle and northern sections of the Longmen Mountain in the upper sub section of the second section of the Maokou Formation are the development areas of high-energy dolomitized platform edge sand debris shoals. The reservoir lithology of the upper sub section of the Maokou Formation is mainly composed of residual bioclastic sandstone dolomite, bioclastic grain dolomite, and residual grain dolomite.The reservoir lithology of the lower sub section of the second section of the Maokou Formation is mainly composed of residual sandstone dolomite and residual grain dolomite. The storage space consists of intergranular dissolution pores, intergranular dissolution pores, intragranular dissolution pores, and expansion fracture pores. The second section of the Maokou Formation is Ⅱ to Ⅲ porous reservoir. The reservoir foundation is composed of high-energy beach deposits. The leaching and dissolution of atmospheric fresh water during the contemporaneous period is the key to reservoir formation. The reservoir has been improved and maintained through shallow burial dolomitization.

In order to avoid wellbore failure caused by abnormal annulus band pressure and resulting safety accidents, the annulus band pressure value is accurately predicted, and preventive measures are taken in advance when it exceeds the control value. An autoregressive integrated moving average-long short term memory (ARMI-LSTM) model was proposed. The model was trained to predict the annular band pressure of example wells based on actual annular band pressure time series data and feature capture data sets, and compared with a single model and recurrent neural network (RNN) model. The results show that the model has a good performance in error, fitting accuracy and overall performance after training with actual data, which can provide a reference for improving the prediction accuracy and efficiency of annular band pressure value, and is helpful to well integrity design.

As an innovative new method of natural gas hydrate extraction, the recovery of hydrate particles determines the efficiency of this method and is one of the key technical links. However, due to the limited hydraulic suction, the recovery rate of hydrate particles is low. Therefore, in order to improve the recovery performance of hydrate particles, a double jet recovery scheme was innovatively proposed. In order to explore the efficiency of solid particle recovery under different working conditions, experimental research was carried out, and the recovery flow field and particle recovery rate were studied by numerical simulation. The results show that with the increase in the distance between the front-end jetting and the recovery hole (l₁) and the distance between the back-end jetting and the recovery hole (l₂), the flow field effect between the double jets weakens, resulting in the particle deposition phenomenon becoming more obvious and the number of sand piles changing from 2 to 3. When the distance between the back-end jetting and the recovery hole (l₂) is 300 mm, the particle recovery increases first and then decreases with the increase in the distance between the front-end jetting and the recovery hole (l₁). When l₁ = 300 mm, the recovery increases first and then decreases with the increase of l₂. The results further enrich the mechanism of hydrate mining and help optimize the design of mining tools.

The role of gas storage in regulating natural gas peaks is crucial. Improper allocation of gas injection schemes during the injection process not only results in excessive energy consumption by compressors, but also leads to excessive pressure changes in certain individual wells and convergence of salt karst cavities, thereby affecting the long-term stable operation of gas storage. By combining the simulated annealing algorithm with actual field conditions, a multi-objective optimization function was established considering both compressor energy consumption and dispersion degree of wellhead pressures across all gas storage wells within the same block. The variable for this optimization was set as the gas injection volume during the task period for each gas storage well, while variables such as maximum design pressure of pipelines, minimum operating pressure, and maximum operating pressure of gas storage wells were taken into account. Additionally, constraints were imposed based on the maximum design flow rate measured by target flowmeters for multi-objective optimization purposes. Results indicate that compressor power consumption can be reduced by over 40% and formation pressure differences can be decreased by more than 90%. It is evident that this scheme provides assurance for ensuring long-term stable operation of gas storage through effective guidance on actual production operations.

In order to better protect the safety of bridge piers and ships, a step-by-step progressive high energy consumption anti-collision magnetorheological damper for bridge piers was designed for the problems of passive energy dissipation, poor dynamic response and anti-collision energy dissipation, and limited adaptability of the device. The structural parameters of the damper were established through the establishment of a mechanical model. The structural strength and magnetic circuit were analyzed using finite elements. The results show that the strength of the structural components meets the requirements, the magnetic induction strength at the gap under 2 A current can reach about 1.2 T, three pistons are set up in the cylinder barrel, which can work together under different crash depth displacements to divide the energy dissipation of the damper into three stages, the minimum damping force under no current is 15 kN, the maximum damping force is 496 kN, and the damping force improves with displacement by 481 kN. The damping force under 2.5 A current increases with displacement from 78 kN to 1 204 kN, which is an improvement of about 15 times. It effectively improve the force of the damper and achieve the excellent effect of graded progressive impact energy dissipation. The application in the bridge pier collision avoidance device can achieve semi-active collision avoidance energy dissipation. The theoretical and finite element simulation results basically coincide with each other, proving the rationality of the damper design.

In order to solve the problems of excessively long droplet break time and excessively large droplet length to diameter ratio in the production of high-temperature and high-viscosity temporary plugging agent, the minimum velocity of droplet forming of temporary plugging agent was obtained by theoretical calculation, and the formula for calculating the optimal disturbance period was derived. The numerical simulation method was used to analyze the droplet forming process of temporary plugging agent, and the changes of the flow field of the droplet forming under the action of no disturbance and external square wave disturbance were explored. The simulation results show that without disturbance, the jet is difficult to break into droplets within 300 mm, the jet temperature is basically unchanged within 0.5 s, and the jet velocity increases to 2.13 times of the initial velocity. When square wave disturbance is added to the outside, too short disturbance period is not conducive to uniform droplet forming, and too long disturbance period will lead to excessive elongation of liquid column before fracture. When the disturbance period is about 0.11 s, the droplet forming efficiency is the highest, the fracture frequency is stable at 0.11 s, and the final droplet length-diameter ratio is stable at about 2. The research results provide a basis for the selection of process parameters for droplet forming of high viscosity temporary plugging agent.

Aiming at the problem of redundancy in the phase space reconstruction of sample entropy algorithm, the phase space reconstruction process of sample entropy algorithm was replaced by a symbolic variable matrix, and an improved sample entropy algorithm was established. The analysis of white noise and powder noise simulation signals shows that the improved sample entropy algorithm can extract signal features effectively and has high computational efficiency. In the past, bearing clearance faults of complex compressors were studied, and the improved sample entropy algorithm was applied to extract features and compared with sample entropy. The feature extraction results of the method are highly consistent with the sample entropy algorithm, and the computational efficiency of the algorithm is much higher than that of the sample entropy algorithm.

In order to explore the application of carbon fiber composite in portable weapons, the technical research on the structural properties of the launcher and the method of composite laminating was carried out. The working pressure of a certain type of launcher was obtained through experiments, and different carbon fiber composite layering models were established by using finite element numerical simulation method. Then the mechanical properties of the launcher under working load were studied. The results show that the layering method of carbon fiber has a great influence on the stiffness and damage failure of the launcher. Under the condition of the same working pressure, the same number of layers and the total thickness, the maximum deformation of the launcher structure is 2.41 mm in the layering mode Ⅰ. The maximun deformation of mode Ⅱ is 7.66 mm, and the latter is more than 3 times of the former. According to the Hashin failure criterion, there is no damage in the carbon fiber cylinder of the mode Ⅰ, except for the 6th to 12th fiber layers in the middle of the cylinder according to the mode Ⅱ, a large range of initial damage points are found in other layers. The research results can provide reference for the stucture design of the launcher and the carbon fiber composite layering scheme.

To enhance the grid connection stability of virtual synchronous machines, a global optimization design method for virtual synchronous machine control parameters was proposed. Firstly, a small-signal model of the virtual synchronous generator with virtual exciter and governor was established, and the system eigenvalues were obtained by solving the state matrix. Secondly, the sensitivity of the controller parameters to the position of eigenvalues was studied, and a wide range of parameter optimization was conducted using genetic algorithms based on the main eigenvalue positions. Finally, analytical solutions of the model and MATLAB/Simulink simulation data were compared. The results show that significant improvements in frequency stability can be achieved by optimizing a wide range of virtual synchronous machine parameters. After optimization, the system response transient stability time is 0.25 s, only 5% of the transient stability time with general parameters, and the frequency stability improves noticeably with load changes.

A novel multi-objective optimization design method for permanent magnet roller based on improved particle swarm optimization algorithm and RMxprt co-simulation was developed to address the issue of low optimization efficiency when relying solely on expert experience. Firstly, an improved particle swarm optimization algorithm was proposed to enhance the convergence speed of the optimization process. Secondly, based on the analysis of the relationship between the structural and performance parameters of the permanent magnet roller, the variable parameters, constraint parameters, and optimization parameters for the improved particle swarm optimization algorithm were determined. Lastly, a MATLAB program for the improved particle swarm optimization algorithm was developed to achieve closed-loop iteration and comparative optimization of the input and output parameters in RMxprt, thereby improving the efficiency and effectiveness of the optimization design for the permanent magnet roller.

With the acceleration of the transition to new energy systems, it is urgent to carry out in-depth research on the complex energy characteristics of multi-load users. A technology of constructing user energy characteristic label library and a user portrait method were proposed, which comprehensively considered the coupling characteristics of electric, cold and thermal multiple loads. Firstly, the high redundancy and low correlation features were eliminated by the fast correlation filtering algorithm, and the features with strong distinguishing ability were selected by the random forest and recursive feature elimination algorithm. In the clustering stage, the improved three-way adaptive density peak clustering (3W-ADPC) algorithm improved the load clustering effect by combining the adaptive neighbor search and the three-branch clustering algorithm. The empirical results show that the proposed method has dual advantages in computational efficiency and clustering accuracy, and can accurately reveal the comprehensive energy use characteristics and deep information of multi-load users, which confirms the practical value of the proposed method in the study of multi-load users’ behavior.

Aiming at the problems of anti-noise, anti-high resistance and complex threshold setting of traditional pole selection methods, a fault selection method of flexible DC distribution line based on Res-BiLSTM network was proposed. Firstly, the original fault signal was subjected to complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), and then the reconstructed signal was obtained by using the correlation coefficient and Shannon entropy for reconstruction. Secondly, the Res-BiLSTM network model was constructed for the pole selection. In order to improve the network accuracy and the convergence speed, the channel attention module was introduced into the split-attention network. The reconstructed signal features were extracted using the convolutional bidirectional long short-term memory and the improved split-attention network at the same time. The extracted features were fused using the attention feature fusion module, and the fused features are classified. Finally, PSCAD/EMTDC was employed to construct the model and to verify the proposed methodology. The simulation results show that the proposed pole selection method is highly accurate, anti-interference, and independent of fault distance.

Under the dual carbon background, a multi-domain communication architecture was constructed to address issues such as a single communication mode and poor adaptability between business and communication technology in the new power load management system. The key supporting technologies of this architecture were comprehensively analyzed. Additionally, an adaptability evaluation system for business and communication technology was developed based on varying business types and their specific communication requirements. A communication technology adaptation method was proposed, employing the fuzzy analytic hierarchy process (FAHP), the CRITIC method, and grey relational analysis-technique for order preference by similarity to an ideal solution (GRA-TOPSIS). The proposed method facilitates the analysis of adaptability between differentiated business requirements and multi-domain communication technologies. The analysis of case studies indicates that the proposed architecture and the adaptation method offer an effective theoretical basis and solution for selecting multi-domain communication technology for the new power load management system business.

Aiming at the problem of white noise amplification of differential microphone array, a design method of parametric differential beamformer was proposed. Through theoretical derivation, it is proved that the delay summing beamformers can maximize the white noise gain and the superdirected beamformers can maximize the directional gain. The orthogonal eigenvector was obtained by using the unitary diagonalization method to deal with the pseudo-correlation matrix between the steering vector and the white noise gain, and the parameters of the beamformer machine were designed based on it. Through simulation experiments, the performance of parametric differential beamformers under different parameter settings was analyzed. Experimental results show that the proposed method can flexibly balance and adjust the white noise gain and directional gain by adjusting the parameters.

Multi-label learning is a common problem in real application scenarios. The construction of large-scale multi-label datasets often means high cost, so semi-supervised learning technology appears. At present, most semi-supervised learning is mainly used in the field of single label classification. Although semi-supervised learning in the field of multiple labels classification has made some progress, there is still much room for improvement in training time consumption, training effects and the use of potential relationships between labels. A multi-label semi-supervised curriculum learning model was proposed under the dual structure semi supervised course learning under dual structure(SSCD) to solve the above problems. Firstly, a curriculum learning scheme based on dual difference was designed, which greatly reduces the training time and improves the robustness of the model. Secondly, a single attention mechanism was designed to explore the potential relevance between labels. The performance of SSCD in the prediction task was evaluated on three open test datasets, and the results compared with four benchmark models show that the comprehensive indicators of SSCD are optimal in all aspects. Finally, the structure ablation experiment was carried out to prove the effectiveness of the proposed single attention mechanism.

In response to the problem of low accuracy in epilepsy detection and recognition using single-view networks, a multi-view convolutional network model with fused attention mechanism (FAM-MCNN) was proposed. Multiple view features were extracted from time domain, frequency domain, time-frequency domain and nonlinear domain to characterize electroencephalogram(EEG) signals comprehensively. Multi-scale convolution was used to capture different levels of detail information. In order to improve the ability to distinguish different types of EEG signals in epileptic patients, the attention mechanism was introduced to combine the features from view dimension and single feature vector dimension respectively. The results of the comparison experiments performed on the CHB-MIT epilepsy dataset show that the average accuracy, sensitivity, and specificity of the FAM-MCNN model are improved by 14.29%, 16.13%, and 12.54%, respectively, when compared to a single-view network. In addition, experiments under a small number of training samples (25%) show that its detection performance reaches the level of the comparison model with a large number of training samples (80%~90%).

Aiming at the problem of time-consuming and labor-intensive routing path design in the cable layout design of complex electromechanical products, an automatic routing technology for complex electromechanical product cables based on multi rules particle swarm algorithm was proposed. Firstly, the cable routing environment of electromechanical products was analyzed, and the routing path was abstracted into a sequence of points to complete the definition of cable routing space. Through pose transformation, the problem of difficult interference detection between wiring paths and parts in electromechanical products was solved. In order to make full use of the wiring space, the particle multiple rules were introduced into the particle swarm optimization algorithm. By using particle number, multi-scale collision detection, adjacent waypoint replacement method and fourth-order quasi-uniform B-spline curve method, the problem that the routing environment is complicated and the optimal solution cannot be obtained was solved, and the searching ability, solving speed and routing quality of the algorithm were improved. Through simulation analysis and comparison with other algorithms, the superiority of the algorithm is proved. The example proves that the proposed method can search feasible paths efficiently during routing. The generated routing paths do not interfere with parts in three-dimensional space, and there are no mutation points in the path fairing, which provides a new idea for the automatic routing of complex electromechanical products.

Compared to image instance segmentation in general scenes, instance segmentation in complex stacked scenes is affected by complex situations such as severe occlusion and stacking of similar objects, making instance segmentation more difficult. To solve the problem of garbage instance segmentation in complex stacking scenarios, an instance segmentation algorithm combining YOLOv8 and two-layer feature network strategy was proposed. Firstly, the feature data was layered in the data preprocessing part, and the two-branch feature fusion was realized through the graph convolutions network (GCN), which reduces the influence of stacking on the features of the occluded objects, thus solving the instance segmentation problem under complex stack occlusion. At the same time, in order to solve the problem that similar objects are easily confused, a soft threshold non-maximum suppression algorithm and a new intersection ratio algorithm were integrated. Finally, according to the complexity of application scenarios and data sets, the feature extraction module of the backbone network was optimized, and the multi-scale attention mechanism was introduced in the backbone network, which effectively improves the detection performance of the model. In the experiment, examples of occlusive garbage classification were used to segment the dataset. The experimental results show that this method outperforms other methods in terms of average accuracy, average accuracy when the intersection to union ratio threshold is 0.5 (AP₅₀), and average accuracy when the intersection to union ratio is 0.5~0.95 (AP_50~95). Compared with the original YOLOv8 algorithm, the detection AP₅₀is increased by 7.9% and the segmentation AP₅₀ is increased by 5.4%, which has better detection and segmentation effects.

In image inpainting, it is crucial that the identification and inpainting of local detail features and the preservation of global features. The models based on fractional-order partial differential equations were characterized by rich evolutionary behaviors, which allow image details to be effectively understood and a certain sharpening effect to be exhibited in image inpainting. However, issues such as inaccurate identification of large-scale features and over-sharpening are prone to be encountered. An optimal control model was proposed and the objective function was defined by the total variation energy of image global features and the constraint was formulated by a spatial fractional-order vector-valued Cahn-Hilliard equation, aiming to achieve a balanced effect between local detail restoration and preservation of global features. L₂ gradient flow, H^-1 gradient flow, and convex splitting were applied to design a numerical scheme for non-convex constraint conditions. And then the split bregman method was used to optimize the objective function with a dynamic grayscale adjustment strategy was introduced to maintain grayscale discrimination capability while enhancing computational efficiency. The numerical experiments demonstrate that the new model achieves an improvement on peak signal to noise ratio(PSNR) ranging from 0.371 8 dB to 9.935 2 dB compared to other methods, exhibiting strong competitiveness in terms of structural similarity(SSIM) and greater effectiveness on images with fragmental damages. Moreover, compared to traditional fractional-order equation models, the computational time is reduced by a factor of 49.50% to 52.91%.

Weld defects present within pipelines constitute a considerable threat for leakage and rupture accidents. To elevate the detection precision of these defects, X-ray inspection was employed as a means to identify and locate them with greater accuracy. However, the diverse types, small sizes, and complex backgrounds of weld defects posed challenges for accurate detection. To address the limitations of current deep learning-based models, such as inadequate adaptability to complex backgrounds and lighting variations, as well as poor performance in detecting small targets, an improved faster region convolutional neural networks(Faster R-CNN) network model was investigated. This model incorporated a channel attention mechanism into the backbone network, modified the residual block structure, and employed ROI Align to replace the traditional ROI Pooling. The results show that compared to the original algorithm, the improved Faster R-CNN model achieves significant improvements in mean average precision (mAP) and F₁, with respective increases of 15.82% and 16.44%. It is concluded that this improved model can meet the high-precision requirements for weld defect detection and holds significant theoretical importance as well as promising prospects for engineering applications.

Aiming at the blindness of the current coal mine roadway surrounding rock support programme and its parameter design, in order to improve the effect of roadway surrounding rock support and meet the requirements of safe and efficient production of mines, taking the Jiaoping mining area as the engineering background, the roof strength, coal gang strength, bottom plate strength, the basic top comes to press the equivalent, mining disturbance, roadway buried depth, roadway protection coal pillar width, span height ratio, top height ratio and the maximum horizontal principal stress were selected as roadway stability master control indicators, and 16 typical roadways and chambers were selected as samples, and the weights of 10 classification indicators were determined based on the analytic hierarchy process. On this basis, the stability of the sample roadway was clustered and analyzed, and the optimal classification number was selected according to the F-statistic method to divide the sample roadway into five categories: very stable, stable, basically stable, unstable and extremely unstable, and then the cluster center of the stability of the mine roadway was constructed. Finally, based on the above theory, the stability of the surrounding rock in the 2407 return wind channel of Yuhua Mine was predicted, and the targeted support measures and parameters were proposed. The results show that the classification results of 2407 return wind channel roadway stability are in line with the actual field engineering, and the deformation control effect of surrounding rock is good, which provides a strong guarantee for the safe and efficient production of the working face.

In September 5, 2022, a M6.8 earthquake occurred in Luding County Sichuan Province. Quite a lot of store-front type buildings damaged or collapsed. Five representative buildings representing both positive and negative aspects were selected to analyze the earthquake damage mechanism through theoretical basis and model experiment. The results show that the earthquake damage is mainly concentrated in the bottom layer, which is composed of concrete column and masonry wall. The masonry wall with no lateral openings restrains the transverse and torsional deformation of the structure, and the floor only transports along the longitudinal direction, and the seismic shear force shared by each member is proportional to its longitudinal lateral stiffness. The rigid and brittle members with large stiffness will appear “internal force condensation”, and then reach “deformation saturation”, and lose the load-bearing capacity in the way of brittle failure, and the gravity of the upper layers will be borne by the transverse wall. If the earthquake does not stop at this time, the transverse wall will lose the role of “buttress” in the direction of exit plane, and the whole structure will collapse along the longitudinal contact with the ground. On the contrary, if the structure of the bottom layer avoids large differences in stiffness, it can significantly improve the seismic resistance.

The drainage consolidation characteristics of municipal sludge are closely related to its water occurrence form. However, there is insufficient understanding of the water transformation law of municipal sludge after chemical conditioning and consolidation. Based on the theory of soil science, the soil-water potential curves of different types of municipal sludge were tested by centrifuge method, and the water forms of municipal sludge were divided into bound water, capillary water and gravitational water according to the range of soil-water potential. On this basis, the different forms of water content in the original sludge, consolidated samples and conditioned sludge samples were compared to reveal the water transformation law of municipal sludge under the action of consolidation and chemical conditioning. The results show that the bound water content of the sludge is reduced by 70%~80% and the free water content is doubled after the sludge is modified by ionic salt. Under the consolidation pressure of 3.1 kPa, only a part of gravity water is discharged from different types of municipal sludge, and the contents of capillary water and bound water are basically unchanged. Under the consolidation pressure of 100 kPa, the gravity water is completely discharged, the capillary water is significantly reduced, and the bound water is slightly reduced.

In order to explore the durability of microbially induced calcite precipitation(MICP) technology to improve the durability of aeolian sandy soil materials, 0.08% polymer absorbent resin (MICP+A) and 0.37% xanthan gum (MICP+B) were used to improve the traditional MICP materials. The microstructure of different cycles of high and low temperature cycle and ultraviolet irradiation was studied by nuclear magnetic resonance technology, and the durability of mineralized sandy soil materials was investigated. The results show that the porosity of both MICP+A and MICP+B materials increases with the increase of cycling cycles. In 20 cycles of high and low temperature cycling tests and 15 cycles of ultraviolet irradiation tests, the MICP+A material shows good stability, and the porosity increment decreases by about 1.8 times and 1.1 times, respectively, compared with the conventional MICP material. Under high and low temperature cycling and ultraviolet irradiation, the crystal structure of calcium carbonate is altered and the percentage of medium-sized pores in the soil increases, causing the 2nd peak of the T₂ spectra of all three materials to be higher than the pre-test peak. The test shows that the polymer water-absorbing resin can improve the stability performance of the traditional MICP specimens, and this study provides a basic experimental basis for the engineering application of microbial mineralized geotechnical materials in the treatment of desert areas.

The low-load one-time construction method of PVC micro-pipe jacking was applied in China and is in its infancy. The small stiffness of the flexible PVC pipe ring leads to the redistribution of soil pressure and the generation of elastic resistance during the jacking process. The jacking force can not be accurately calculated by the existing specifications. The principle and construction method of the PVC micro-pipe jacking method were used to investigate. At the same time, the influence of vertical deformation of PVC pipe on soil arching effect and the additional frictional resistance caused by horizontal deformation were considered based on Terzaghi soil pressure. The corresponding jacking force calculation formula was derived on this basis. Furthermore, the factors that influence the jacking force of PVC pipes were analyzed through finite element numerical simulation. The results indicate that the pressure around the PVC pipe will increase due to the deformation of the pipe. The maximum error between the predicted lower limit and the measured value of the jacking force calculation model proposed is 15%. The measured value is between the predicted upper limit and the lower limit, which proves its applicability. The parametric analysis indicates that the jacking force increases by 1.6 times as the pipe diameter increases by 100 mm and the jacking force increases by about 3 times when the buried depth of the pipeline increases from 4.5 m to 6 m.

In the condition of engineering surcharge near the existing pipeline, the pipeline produce subsidence deformation and further threaten the normal operation of the existing pipeline. Most of the research in this area stays in the finite element and indoor tests, and few theoretical solutions are used to analyze the stress and deformation response of existing pipelines under adjacent engineering surcharge. Based on this, the pipeline-soil interaction under this working condition was investigated by using theoretical analysis. Firstly, the Boussinesq solution was used to analyze the additional stress at the axis of the existing pipeline. Then, the pipeline was simplified as an infinite beam rest on the Pasternak model to further obtain the total energy of the system during the deformation of the pipeline. Finally, the stress and deformation response of the pipeline can be obtained according to the energy variation theory. By comparing with the existing experimental data, the correctness of the proposed method was verified. Compared with the degradation analysis of the proposed method, the proposed method is closer to the measurements. The parameter study shows that the stress deformation of the pipeline wills decrease nonlinearly with the increase of the buried depth of the pipeline. Increasing the diameter of the pipeline would increase the deformation response of the pipeline, the deformation of the pipeline is not sensitive to the angle between the pipeline and the loading area. Increasing the horizontal distance between the pipeline and the loading area can effectively reduce the stress and deformation response of the pipeline, and the deceleration increases first and then decreases. A series of analysis results can be used to analyze the influence of engineering surcharge on the stress and deformation of existing pipelines in practical engineering.

In order to study the influence of slurry shield tunneling parameters on surface settlement, based on the slurry shield tunneling and monitoring data of the left line of the Hesong-Heshan stacked section of Harbin Metro Line 3 project, based on the BP neural network optimized by genetic algorithm, the different settlement output forms were studied. The tunnel distance label was introduced to optimize the neural network fitting effect, and the parameter sensitivity analysis was carried out according to this network model. Three most sensitive parameters were obtained, and exhaustive tests were carried out to further analyze the specific influence of parameters on surface settlement. The research shows that the surface settlement performance of slurry shield tunneling is not closely related to the tunneling parameters after passing through a certain ring for two days, and the surface settlement analysis can focus on the monitoring value of the day. Before, during and after the shield machine passes through a certain ring, it will have different effects on the surface settlement above the ring. Subsequent research on surface settlement based on neural network can be considered to include this index. Among the parameters of slurry shield tunneling, reducing slurry viscosity and increasing slurry specific gravity can control surface subsidence, and increasing propulsion speed can reduce the impact of construction on surface subsidence.

The construction of the shield tunnel for the high-speed railway will impact the settlement of railway tracks, roadbeds, and other structures, potentially affecting the railway’s operation. In order to study the influence of shield tunnel underneath the high-speed railroad in Guiyang area on the settlement pattern of frame box culvert, roadbed and high-speed railway track and the shield construction parameters suitable for Guiyang area, relying on Guiyang Rail Transit Line 3, the model of soil layer-frame culvert-roadbed-track was established to analyze the settlement law of shield construction in limestone on frame culvert, existing high-speed railroad track and roadbed, and to optimize the parameters of shield construction. The results indicate that the settlement of the frame culvert exhibits a “W” shaped transverse and longitudinal settlement pattern, with a peak settlement value of 1.805 mm during the tunneling process. The shield machine’s crossing causes a sudden change in subgrade settlement, with the peak value reaching 1.753 mm. High-speed rail track settlement shows a single peak with excavation, with a peak value of 1.41 mm. Analysis of eight types of shield tunneling pressure and grouting pressure reveals that grouting pressure has a better control effect on settlement than excavation pressure. Increasing grouting pressure leads to a decrease in the peak value of subgrade surface settlement, with the peak value reaching 1.355 mm at 800 kPa grouting pressure. The maximum deformation rate is 20.7%, with grouting pressure set at 800 kPa and tunneling pressure at 100 kPa for the shield machine.

Rockburst is an extremely destructive geological disaster in deep underground engineering. In order to accurately predict the intensity level of rockburst, a method for rockburst intensity level prediction based on parallel fusion graph Transformer (PFGT) was proposed. Firstly, the similarity structure relationship of rockburst data in Euclidean space was utilized to construct graph-structured data. Besides, another kind of graph-structured data was constructed by utilizing multiple rockburst criteria to constrain the structural distortion of rockburst data in European space. Single-scale features of rockburst data was obtained through parallel training. Secondly, a feature fusion graph Transformer strategy was designed, which obtains multi-scale features of rockburst data by fusing two types of graph-structured data features based on Euclidean space and based on rockburst criteria. The method improves the data representation capability by simultaneously utilizing single-scale features and multi-scale features. During the training process, using Transformer for feature fusion enables the model to more comprehensively capture the optimized features of rockburst data, thus improving model performance. Compared with traditional neural networks and other machine learning algorithms, the prediction accuracy of the PFGT model is 94.87%, which is superior to other algorithms, proving the effectiveness of this algorithm and providing a new method for rockburst level prediction.

Restricted by topographic conditions and surrounding environment, the design spacing of double-hole tunnels in western China is often close. Therefore, the special structural form of multi-arch or small close tunnels is often used in the tunnel design. However, in the western region of China, high-intensity seismic zones are widely distributed, and the middle rock wall between the close tunnels is more susceptible under the strong earthquakes, especially the near-field ground motions, which results in large plastic deformation of local surrounding rock, and the safety operation of the tunnel was affected. Therefore, in order to study the distribution of plastic zone of surrounding rock of small close-distance tunnel under near-fault ground motion, reasonable near-fault ground motions were selected, and 3D finite element numerical models considering different tunnel close-distance and surrounding rock conditions were established. The dynamic response characteristics of small close-distance tunnel and the distribution law of plastic zone of surrounding rock were revealed. Thus, through the grouting reinforcement of the middle rock wall, the seismic effectiveness of the reinforcement measures on the plastic zone was verified. The results show that the acceleration response of the arch foot near the middle rock wall tunnel is more significant under the influence of near-field ground motion. Under transverse seismic excitation, the maximum principal stress value of the left spandrel and the right arch foot of the right line tunnel are more significant, while the minimum principal stress of the side walls and the inverted arch of the tunnel is larger. The plastic zone of surrounding rock of small-close-distance tunnel under strong earthquake is greatly affected by rock mass grade. When the tunnel spacing is 0.50 times of the tunnel span, the development of plastic zone in grade Ⅳ and grade Ⅴ surrounding rock is the most serious, and the phenomenon of plastic zone penetration of surrounding rock at the haunch of tunnel near the middle rock wall appears. Taking the plastic zone of surrounding rock as the discriminant index, the grouting reinforcement of middle rock wall is beneficial to the seismic performance of small close distance tunnel. The research results can provide a reference for the seismic safety of small close-distance tunnels in high-intensity earthquake areas.

Accurately predicting bike-sharing flow is essential for optimizing the supply-demand balance of shared bikes and enhancing urban residents’ travel convenience. To address the issues of low prediction accuracy and insufficient capture of spatiotemporal characteristics in bike-sharing flow prediction, a hybrid convolutional-recurrent neural network (Conv3D-GRU) model was proposed. Using Chicago’s 2022 full-year bike-sharing data, experiments were conducted, and the results were compared with those of the 3D convolutional neural network (3D-CNN) model and the convolutional long short-term memory (ConvLSTM) model. The model performance was evaluated using root mean squared error (RMSE), mean absolute error (MAE), and the coefficient of determination (R²). Experimental results show that compared with the 3D-CNN and ConvLSTM models, Conv3D-GRU is improved by 3.25%, 4.90%, 1.14% and 11.94%, 13.70% and 2.46% on RMSE, MAE and R², respectively. This demonstrates that the Conv3D-GRU model has lower prediction errors and higher prediction accuracy, making it an effective and reliable approach for forecasting bike-sharing inflow and outflow.

In the absence of accurate transit demand information, a demand responsive transit(DRT) route planning method based on taxi trajectory data was proposed to predict the “potential demand” of demand responsive transit and provide a feasible plan for route planning before transit operation. Firstly, taxi trajectory data in the study area was obtained through data mining, representing the “potential demand” for passenger travel in the area, and candidate station were determined using the K-means clustering algorithm. Secondly, a benchmark station network was established using these candidate station, with edge benchmark stations designated as the starting and ending points of routes. Utilizing the K-shortest pathes(KSP) algorithm constrained by route length, benchmark chains were generated. Finally, after determining the sub-chain set of the benchmark chains, demand response stations within each sub-chain were searched based on circumferential critical value constraints. Using this algorithm, alternative routes were generated repeatedly within specific time periods, and an initial optimal route was selected based on comprehensive evaluation indices for each alternative route.

A numerical calculation model for airfoil dynamic stall numerical simulation was established by computational fluid dynamics method, and the influence of plunging motion on airfoil unsteady aerodynamic force was analyzed. The simulation results were compared to the pitching motion wind tunnel experimental data of NACA0012 airfoil, and the results of the model under mild stall and deep stall conditions were in good agreement with the experimental values, which verified the accuracy and feasibility of the numerical calculation model. The plunging motion of NACA23012 airfoil was equivalent to the pitching motion, the lift characteristics of airfoil under the two motion modes are very close, but the moment characteristics are obviously different. With the increase of amplitude of plunging motion and inflow Mach number, the difference of moment characteristics is further expanded. With the increase of amplitude of plunging motion, the damping effect of aerodynamic moment of pitching and plunging motions is obviously enhanced, with the increase of inflow Mach number, the damping effect is reduced, and the moment divergence occurs when the Mach number is 0.85.

The experimental research on the dynamic response of aero-engine rotor under sudden impact load was carried out by using high-speed motor drive on the vibration table. Rotor dynamic response tests under different characteristic speeds, load sizes, impact directions, and pulse widths were completed, revealing the general rules of rotor dynamic response under sudden impact loads. The results show that the dynamic response of the rotor increases instantly and then returns to a stable state when under instantaneous impact, and it increases with the increase of the sudden impact load. Besides, the vertical dynamic response of the rotor is greater than the horizontal dynamic response, and the vertical response of the same measurement section is 4% to 46.15% greater than the horizontal response under various operating conditions. In addition, the dynamic response of the rotor under axial foundation impact is greater than that under vertical foundation impact. Within a certain range, as the pulse width of the impact load increases from 6 ms to 11 ms, the dynamic response of the rotor decreases by 2.5%~10%. The study provides a reference for the vibration response analysis of aero-engine rotors under sudden impact loads and the structural safety design of aero-engine, which has important engineering application value.

With the development of global trade, buoyancy-lifting hybrid airship is an important choice for global long-distance and large-load transportation, it has gradually become a research hotspot at home and abroad. In order to improve the transportation efficiency of buoyancy-lifting hybrid airship, its layout and parameter sensitivity were studied. Discussions were conducted on the hull layout and tail wing layout, and several types of hull and tail wing layouts with high lift to drag ratios were proposed. At the same time, a sensitivity analysis was conducted on the impact of design parameters on aerodynamic characteristics. The results indicate that the width of the hull has the most significant impact on the lift coefficient, drag coefficient, and maximum lift to drag ratio of airships. Its relative sensitivity coefficient is nearly ten times that of the longitudinal position and inclination angle of the tail wing. The inclination angle of the tail wing has the strongest impact on the torque coefficient of the airship, and its relative sensitivity coefficient is more than twice the width of the hull.

With its advantages in cooling efficiency and cost, the cold channel closed system is more and more used in the construction of new data centers. However, the closed channel system will turn the originally open channels between data center cabinets into narrow spaces with restricted ventilation. When the main combustibles in the data center catch fire, the accumulated hot smoke and gas in the closed channel cannot be timely discharged, seriously threatening the safety of data center equipment and personnel. At present, there is a paucity of experimental data and theoretical basis for the fire hazard of cables commonly used in data centres. Pyrosim software was used to establish a full-size physical model of the data center room in the cold closed channel, and fire dynamics simulator(FDS) software was used to establish a full-size fire model to simulate different fire source locations, so as to analyze the changes of fire parameters such as smoke spread rate, visibility and temperature distribution. The results show that when a fire occurs under the floor (maximum heat release rate reaches 2 000 kW), the smoke would fill the whole machine room more quickly due to the influence of the special air conditioning airflow and floor than the cabinet fire. The visibility at the safety exit measuring point reached 0 m, 60 s earlier than the fire at the inside of the cabinet. At the same time, affected by the air conditioning airflow and perforated tiles, the temperature at the inside of the cabinet quickly reaches the critical value where the fire hazard is much greater. The results could provide important theoretical support for the fire protection system design of the cold channel of closed data center.

Process production safety monitoring is the main technical method for safety risk control and accident prevention, and data is a significant basis for safety control and decision-making. In the existing security monitoring network, there are many sensor nodes and large amounts of data, which cause heavy channel load in the wireless sensor network. This often leads to issues such as data latency and loss, which affects the timeliness and accuracy of safety control decisions. Therefore, the security risk factors of typical process production scenarios were focused on. Based on this, the sensor deployment plan and wireless sensor network data transmission architecture were clarified, a security monitoring data flow scheduling method based on superior-subordinate server was proposed, and the data stream congestion index and abnormal data packet frequency index were used as the main indicators of data flow performance evaluation. Subsequently, the chemical polymerization reactor were taken as an engineering scenario, the performance improvement was examined after the subordinate server was initiated to share the data flow for the superior server. Through the comprehensive study of channel load balancing, the method of superior-subordinate server is benefit to ensure the effectiveness of orderly transmission of safety monitoring data and risk control.

In order to investigate the potential causative factors and mechanisms of accidents in the flight transit security system, and to further ensure the safety of civil aviation operation, based on the system theory and the gray correlation theory, and combining with the actual situation of the flight transit security operation process, the safety problems in the system were transformed into the control and feedback problems, and the safety control and feedback structure was mapped out. Using complex network theory to transform accident causation and its logical relationship, a directed weighted accident causation network model was constructed, the overall characteristics of the network and the connection of each node from different perspectives were quantitatively analyzed, such as the node degree, the network diameter and the average path length, etc., and then 16 important accident causation factors affecting the flight transit security system were selected. Through grey correlation analysis, the influence degree of each cause factor on the accident was judged, and the key cause factors that need to be prevented and controlled were finally determined. The results show that the personnel factor dominates the accidents in the flight crossing security system, and its sub-factors, such as speeding, insufficient number of personnel, error of towing personnel and illegal entry of personnel into the control area, are the key causal factors leading to the accidents.