Table of Content

25 May 2025, Volume 53 Issue 5

Previous Issue    Next Issue

Contents



2025, 53(5):  0. 

Asbtract ( 37 )   PDF (429KB) ( 10 )  

Related Articles | Metrics

Mechanical Engineering

Hybrid Toolpath Planning and Feed Rate Optimization for Cavity Machining Under Constant Loads



WANG Qinghui, WANG Jinqiang, DING Xuesong, LIAO Zhaoyang

2025, 53(5):  1-10.  doi:10.12141/j.issn.1000-565X.240467

Asbtract ( 196 )   HTML ( 8)   PDF (4937KB) ( 247 )  

Figures and Tables | References | Related Articles | Metrics

The CNC machining of molds and various 3D parts involves numerous cavity features, and the design of machining toolpaths directly affects machining quality and efficiency. With the advancements in high-speed milling technology, CNC machines provide the hardware foundation for improving cavity machining efficiency, but they also place higher demands on CAM toolpath design. Traditional CAM toolpaths tend to cause abrupt changes in cutting load when dealing with areas such as cavity corners, slots, and intersections of circular paths. This load instability limits the improvement of feed rate and cutting depth, negatively impacting both machining efficiency and quality. To address these issues, this paper proposes a hybrid toolpath planning and feed rate optimization method aiming at achieving constant load machining for cavities. The method, which is based on a multi-level block structure, first calculates the material removal rate and then divides the machining areas into stable, semi-stable and load fluctuation regions. For different regions, circular toolpaths, feed speed optimization and variable-radius trochoidal paths are comprehensively adopted to ensure smooth load control throughout the machining process. By applying trochoidal paths in areas prone to load fluctuations, sudden load variations can be reduced and stable machining process can be ensured. Experimental results show that the proposed toolpath planning and feed rate optimization method is suitable for generating CAM toolpaths for various complex cavities, with good load stability and better machining quality.

A Quick Coupling Device with Free Posture Adjustment Used for Emergency Rescue Robot Attachments



LI Wei, LIU Jiachen, ZHANG Weiyuan, HUANG Rihong, BAI Jing, JIANG Chao

2025, 53(5):  11-19.  doi:10.12141/j.issn.1000-565X.240408

Asbtract ( 191 )   HTML ( 4)   PDF (4923KB) ( 61 )  

Figures and Tables | References | Related Articles | Metrics

Due to its single function and low flexibility, the existing emergency rescue equipment is difficult to meet the requirements of complex emergency operations under geological disasters such as earthquake. To solve this problem, an electromechanical hydraulic quick coupling device, which can realize fast change of attachments and free adjustment of posture, is designed. The device can be quickly integrated into emergency rescue equipment to complete high-mobility and multi-function rescue tasks. In the investigation, first, the extreme load characteristics, as well as the stress and strain situations during the operation of the device are simulated and analyzed, through which the weak part and load spectrum are determined. Then, the reliability theoretical models considering cyclic damage strength degradation under deterministic and random periodic stress are deduced and established, and the mapping relationship between the reliability and the failure rate of the rotating mechanism is determined. Furthermore, based on linear elastic fracture mechanics, the crack propagation of the weak part of inclined oil cylinder’s piston rod is analyzed, and the fatigue life of the device is determined via the local stress strain method to ensure that the device can meet the application requirements. Finally, the developed device is integrated into the walking rescue robot. Test results show that the proposed quick coupling device can realize the rapid switching of various attachments such as bucket and gripper, with a switching time of less than 15 s, as well as the increases by ±40° yaw and ±360° rotation degrees of freedom, thus meeting the requirements of flexible operation. This research provides theoretical reference for the design of similar devices.

Machining Feature Recognition Method of B-Rep Model Based on Graph Neural Network



HU Guanghua, DAI Zhigang, WANG Qinghui

2025, 53(5):  20-31.  doi:10.12141/j.issn.1000-565X.240329

Asbtract ( 331 )   HTML ( 10)   PDF (2520KB) ( 65 )  

Figures and Tables | References | Related Articles | Metrics

Automatic feature recognition is one of the key technologies of intelligent manufacturing. Traditional rule-based recognition algorithms have poor scalability, and the methods based on deep convolutional networks are of low accuracy because they use discrete models as input and the recognition results are difficult to accurately map back to the original CAD model, causing inconvenience in application. In view of these shortcomings, a feature recognition method based on graph neural network, which can directly analyze B-Rep models, is proposed. The method extracts effective characteristic information and geometric information from the B-Rep structures to form a feature descriptor, and then establishes an adjacency graph with high-level semantic information based on the topological structure of the CAD model. By taking the adjacency graph as the input, an efficient graph neural network model is constructed. By introducing a differentiable generalized message aggregation function and a residual connection mechanism, the model possesses stronger information aggregation performance and multi-level feature capture capabilities. What is more, message normalization strategy is used to ensure the stability of the training process and to accelerate the convergence of the model. After the training, the network can directly classify and annotate all faces in the B-Rep model, thereby realizing feature recognition. Experimental results on the public dataset MFCAD++ demonstrate that the proposed method achieves an accuracy of 99.53% and an average intersection-over-union ratio of 99.15%, which outperforms other similar studies. Further evaluations using more complex testing cases and typical CAD cases from real engineering applications show that the proposed method is of better generalization ability and adaptability.

Mechanism of Saw Tooth-Shaped Chip Formation in Cast & Wrought High-Temperature Alloy GH4198



JIN Qichao, LI Jun, WANG Liangliang, TAN Haibing, LI Fulin, FU Rui, MENG Lingchao, GUO Lei

2025, 53(5):  32-44.  doi:10.12141/j.issn.1000-565X.240390

Asbtract ( 134 )   HTML ( 5)   PDF (6185KB) ( 57 )  

Figures and Tables | References | Related Articles | Metrics

In order to reveal the formation mechanism of saw tooth-shaped chips in cast & wrought high-temperature alloy GH4198 and predict chip morphology through theoretical models, orthogonal cutting experiments were conducted. Based on the slip line field model, the geometric shape of the chips was predicted, and the influence of cutting parameters on chip formation was analyzed. A three-stage formation model of saw tooth-shaped chips considering the tool edge radius was proposed, and a two-dimension orthogonal cutting thermo-mechanical coupled finite element model was established, with its rationality being verified through experiments. By analyzing the variations of stress, equivalent plastic strain and temperature during the chip formation obtained from simulations, the formation mechanism of saw tooth-shaped chips was investigated. The results show that the shear angle increases with the increase in cutting speed and feed rate, while the chip thickness decreases with the increase in cutting speed. At the cutting speeds of 10, 20 and 30 m/min, the relative errors of the predicted chip thickness are respectively 4.20%, 12.34% and 24.73%, the maximum chip thickness compression ratios are respectively 3.19, 2.78 and 2.26, and the chip serration degrees are respectively 0.20, 0.36 and 0.58. At a cutting speed of 30 m/min, obvious cracks appear in the chips, and the saw teeth exhibit an overall inclined shape. At the feed rates of 0.05, 0.10 and 0.15 mm/r, the relative errors of the predicted minimum chip thickness are respectively 17.66%, 8.66% and 5.07%, the maximum chip thickness compression ratios are respectively 2.82, 2.78 and 2.61, and the chip serration degrees are respectively 0.12, 0.36 and 0.42. The slip line field model effectively predicts the variation of chip thickness with cutting parameters. With the increase in cutting speed and feed rate, the chip thickness compression ratio shows a decreasing trend, while the serration degree increases with a gradually slowing trend. Additionally, the influence of the tool edge radius on chip formation was analyzed through finite element simulation, and the effectiveness of the theoretical model for saw tooth-shaped chip formation was verified.

Analysis of Extrusion Law of Large-Scale and Small-Scale Aluminum Profiles with Hollow Thin Wall for Rails



LIU Guoyong, GAO Shize, ZHU Dongmei

2025, 53(5):  45-55.  doi:10.12141/j.issn.1000-565X.240397

Asbtract ( 131 )   HTML ( 4)   PDF (3507KB) ( 19 )  

Figures and Tables | References | Related Articles | Metrics

In order to explore the extrusion law of large-scale and small-scale hollow thin-wall aluminum profiles for rails, simulation software HyperXtrude is used to numerically simulate the extrusion process of the profiles, the influences of mold structure and process parameters on the extrusion are analyzed, and the forming rules of two large-scale and small-scale profiles with similar shapes are compared. The results show that, in terms of mold structure, the modification of welding chamber and drainage groove has the most obvious influence on large-scale and small-scale profiles, for instance, the change in welding chamber significantly reduce the maximum deformation of the small-scale profile, with a reduction of 42.82%, while that for the large-scale profile is 25.34%. The change in drainage groove structure shows different impact trends—after altering the drainage groove, the maximum deformation reduction of the large-scale profile is 40.88%, while that of the small-scale profile is 24.72%. The drainage groove of small-scale profile is relatively shorter, and the modification of the drainage groove of large-scale profile is more complicated, so that the change of drainage groove has a more significant impact on large-scale profile. Moreover, in terms of process parameters, according to the changes of metal deformation, metal flow rate and the SDV values of the profile exit section under different conditions, it is found that the extrusion speed and the die temperature have more significant impact on the large-scale profile, while the billet diameter has a more pronounced effect on the small-scale profile. This research provides theoretical support for optimizing the extrusion process of aluminum profiles.

Computer Science & Technology

A Method for Software Vulnerability Detection via Path Representations and Pretrained Model



LU Lu, WAN Tong

2025, 53(5):  56-65.  doi:10.12141/j.issn.1000-565X.240324

Asbtract ( 181 )   HTML ( 6)   PDF (1609KB) ( 37 )  

Figures and Tables | References | Related Articles | Metrics

Software vulnerabilities are critical weaknesses that compromise the security of computer systems, making them susceptible to attacks may lead to data breaches, system crashes or even more severe security incidents. Therefore, accurately and efficiently detecting software vulnerabilities has become a central research focus in the field of computer security. Although contemporary deep learning-based vulnerability detection approaches have made progress, they are often limited by single code representations and fail to fully capture the complementary nature of code semantics and structural information. This research introduces an innovative method for software vulnerability detection, termed VDPPM (Vulnerability Detection via Path Representations and Pretrained Model), which effectively enhances code semantic analysis and vulnerability detection accuracy. VDPPM integrates the path representations extracted from abstract syntax tree, control flow graph and program dependency graphs, leverages the SimCodeBERT model optimized through contrastive learning framework SimCSE to enhance the model’s ability to capture vulnerability features. In the experiments, first, three types of code representations are extracted from the source code and are used to construct a corpus by deriving path representations for the training of Doc2vec model, thus generating general-purpose embedding models, converting path sequences into vector representations. Subsequently, a pretrained CodeBERT model is integrated, which, after being trained under the contrastive learning framework, gains increased precision in capturing deep semantic features within the code. Finally, by combining vector embeddings from Doc2vec and SimCodeBERT, high-quality code representations are constructed to perform vulnerability detection. Experimental results demonstrate that, across multiple publicly available benchmark datasets for vulnerability detection tasks, VDPPM outperforms the existing mainstream methods with significant improvements in several performance metrics. This convincingly validates the effectiveness and superiority of the proposed method.

Carbon Emission Prediction in Transportation Industry Based on SD-ISSA-DALSTM



WANG Qingrong, WANG Junjie, ZHU Changfeng, HAO Fule

2025, 53(5):  66-81.  doi:10.12141/j.issn.1000-565X.240356

Asbtract ( 281 )   HTML ( 6)   PDF (3678KB) ( 32 )  

Figures and Tables | References | Related Articles | Metrics

Aiming at the low accuracy of carbon emission prediction caused by the high volatility and nonlinearity of the carbon emission data series in transportation industry, a transportation carbon emission prediction model combining the secondary decomposition, dual attention mechanism, improved sparrow search algorithm (ISSA) and long short-term memory (LSTM) network is proposed. First, complete ensemble empirical mode decomposition with adaptive noise is introduced to decompose the transportation carbon emission data series into modal components with different frequencies, then sample entropy is used to quantify the complexity of each component, and secondary decomposition is performed on the component with the highest entropy value via variational mode decomposition, which further weakens the volatility and nonlinearity of the transportation carbon emission data series. Next, in order to explore the correlation between transportation carbon emission and its influencing factors, a double attention mechanism-optimized LSTM (DALSTM) model is constructed, in which a feature attention mechanism is added to the input side of the LSTM to highlight the key input features. Meanwhile, a temporal attention mechanism is added to the output side to extract the key historical moments. Finally, the SSA algorithm is improved by combining the Circle chaotic mapping, the dynamic inertia weight factor and the mixed variance operator strategies, ISSA-DALSTM models are established for each component separately, and the predicted values of each component are reconstructed. By measuring the carbon emission data of China’s transportation industry from 1990 to 2019, it is found that the root mean square error, mean square error, and mean absolute percentage error of the proposed model are respectively 5.308 8, 3.566 1 and 0.443 9, which are better than those of other comparative models, thus verifying the validity of the proposed model.

A Spatiotemporal Heterogeneous Two-Stage Fusion Network for Traffic Flow Prediction



HOU Yue, YIN Jie, ZHANG Zhihao, LU Keke

2025, 53(5):  82-93.  doi:10.12141/j.issn.1000-565X.240480

Asbtract ( 240 )   HTML ( 5)   PDF (3410KB) ( 36 )  

Figures and Tables | References | Related Articles | Metrics

In response to the existing traffic flow prediction studies that fail to fully integrate complex spatiotemporal correlations and heterogeneities, this paper designs a traffic flow prediction network based on grid data, namely the spatiotemporal heterogeneous two-stage fusion neural network marked as ST_HTFNN. This network employs a phased and hierarchical spatiotemporal feature extraction architecture, and adopts a new model where the static and dynamic feature extraction stages are serialized. In the static feature extraction stage, a novel Mamba-like linear attention (MLLA) block is introduced as a static heterogeneous fusion unit to achieve spatial correlation and heterogeneity fusion mining. In the dynamic feature extraction stage, a simple and efficient dynamic heterogeneous fusion unit is designed, and dilated convolution is combined with gating mechanisms to adaptively fuse and capture global and local spatiotemporal correlations and heterogeneities. Furthermore, to address the smoothing of road features during the deep convolution process for road-level traffic flow characteristics, a road feature enhancement module is designed to reconstruct and enhance road information. Finally, an external disturbance feature fusion module is designed to integrate the impact of external disturbance features on traffic flow prediction results. Experimental results on three real-world traffic datasets, namely BikeNYC, TaxiCQ and TaxiBJ, demonstrate that the ST_HTFNN model outperforms the existing benchmark methods, respectively with a decrease of 6.13%, 0.8% and 7.01% in the mean absolute error of prediction accuracy.

A Liver Tumor Image Segmentation Method Based on Multi-Scale Feature Fusion and Reconstruction Convolution



MA Jinlin, JIU Zhiqing, MA Ziping, XIA Mingge, ZHANG Kai, CHENG Yexia, MA Ruishi

2025, 53(5):  94-108.  doi:10.12141/j.issn.1000-565X.240439

Asbtract ( 227 )   HTML ( 7)   PDF (7299KB) ( 42 )  

Figures and Tables | References | Related Articles | Metrics

Aiming at the problem of insufficient expression ability of liver tumor image features and limited global contextual information transmission, an improved U-Net liver tumor image segmentation method is proposed. Firstly, a low-rank reconstruction convolution is designed to optimize the large number of parameter problems caused by traditional convolution operations, and is used to construct a convolution kernel reconstruction module that uses residual structure to improve the encoder decoder, so that the encoder retains more detailed information and the decoder recovers information more effectively, thereby enhancing the expression ability of liver tumor image features. Then, to enrich the transmission of global contextual information, a three-branch spatial pyramid pooling module is designed to optimize the bottleneck structure of information transmission and to break the limitation of a single path. Secondly, a multi-scale feature fusion module is designed to optimize the reuse mechanism of encoder information, enhance the modeling ability of the model for global contextual information, and improve its efficiency in extracting liver tumor image features in different scales. Finally, the performance of the proposed method is tested on LiTS2017 and 3DIRCADb datasets. Experimental results show that the method achieves a Dice coefficient and an IoU value of 97.56% and 95.25% in the liver image segmentation task on LiTS2017 dataset, and 89.71% and 81.58% in the liver tumor image segmentation task. Moreover, the Dice coefficient and IoU value in the liver image segmentation task on 3DIRCADb dataset respectively reach 97.63% and 95.39%, while respectively reach 89.62% and 81.63% in the liver tumor image segmentation task.

Prediction Method of IL-6 Inducing Peptides Based on Graph Neural Network



CAO Ruifen, HU Weiling, LI Qiangsheng, BIN Yannan, ZHENG Chunhou

2025, 53(5):  109-117.  doi:10.12141/j.issn.1000-565X.240242

Asbtract ( 133 )   HTML ( 6)   PDF (1821KB) ( 65 )  

Figures and Tables | References | Related Articles | Metrics

Interleukin-6 (IL-6) is a highly multifunctional glycoprotein factor that can regulate both innate and adaptive immunity as well as various aspects of metabolism, including glycolysis, fatty acid oxidation and oxidative phosphorylation. Many studies have shown that the expression and release of IL-6 in patients infected with viruses significantly increase, and are positively correlated with the severity of the disease. Therefore, identifying IL-6 inducing peptides and exploring their action mechanisms are very important for developing immune therapies and dia-gnostic biomarker for the severity of diseases. Currently, the identification methods of IL-6 inducing peptides mostly use traditional machine learning, in which feature selection and extraction are rather complex, and field expert knowledge are required. In view of this problem, this paper proposes a novel graph neural network-based prediction method of IL-6 inducing peptides named SFGNN-IL6. In this method, the predicted structural characteristics of IL-6 inducing peptides are used to construct the adjacency matrix by screening the distance information according to the threshold, and the node features of amino acids are extracted using One-hot encoding, position encoding and BLOSUM62 encoding, and are then graph-represented. Moreover, graph attention mechanism layers and graph convolutional neural network layers are used as dual channels to separately extract features, considering both the update of node weights and the update of node information. Finally, the two types of features are fused for the classification of IL-6 inducing peptides. Experimental results validate that the proposed method is effective.

Energy, Power & Electrical Engineering

Quantitative Analysis of Interaction in Sub/Super-Synchronous Oscillation of Double-Fed Wind Farm Integrated into Grid Through MMC-HVDC System



ZHU Lin, ZHAO Xinyue, ZHONG Danting, WU Zhigang, GUAN Lin

2025, 53(5):  118-129.  doi:10.12141/j.issn.1000-565X.240078

Asbtract ( 139 )   HTML ( 2)   PDF (2966KB) ( 25 )  

Figures and Tables | References | Related Articles | Metrics

To quantitatively investigate the impact of interactions among control loops in a double-fed wind farm integrated with the grid via an MMC-HVDC system on sub-/super-synchronous oscillations, an analytical method combining modal analysis and relative gain array (RGA) is proposed. Firstly, a small-signal model of the double-fed wind farm integrated with the grid through MMC-HVDC is established, with its accuracy being verified by comparing its step response with that of an electromagnetic transient simulation model. Secondly, modal analysis is employed to identify the dominant sub-/super-synchronous oscillation modes affecting system stability, and the primary participating variables of these oscillation modes are determined through participation factor calculations, laying a foundation for subsequent analysis of the influence of interactions among different control loops. Thirdly, the RGA is introduced to confirm the existence of interactions, quantify and compare the strength of interactions among control loops associated with the primary variables of the dominant oscillation modes. This focuses subsequent research on the rotor-side converter (RSC) control loop of the wind farm and the fixed V/f control loop in the MMC-HVDC system. Finally, based on the variation of RGA values with influencing factors, the effects of the electrical distance of grid connection and the controller parameters on the degree of interaction among control loops are quantitatively evaluated and verified using time-domain simulation. The study reveals that, when the electrical distance increases or the proportional coefficient of the fixed V/f control on the MMC-HVDC side rises, the interaction between the RSC control loop on the double-fed wind turbine side and the fixed V/f control loop intensifies, leading to a decrease in system stability.

Motion Distribution Characteristics of Regular Tetrahedral Particle Population in Vertical Uppipe



PENG Deqi, ZHOU Jingqiang, FENG Yuan, HUANG Zhizhong, TAN Zhuowei, TANG Mingcheng, PENG Jianguo, CHEN Ying

2025, 53(5):  130-138.  doi:10.12141/j.issn.1000-565X.240222

Asbtract ( 117 )   HTML ( 3)   PDF (2959KB) ( 94 )  

Figures and Tables | References | Related Articles | Metrics

Liquid-solid two-phase flow technology is widely applied to the heat transfer enhancement in heat exchanger design, the key lies in guiding low-volume-fraction particles to the wall region to disrupt the thermal boundary layer and thereby improve the heat transfer efficiency. The movement behavior of particles is a key factor for deep analysis of heat transfer enhancement mechanism. Non-spherical particles have better disturbance effects and more complex movement behaviors due to the anisotropy of their shapes. This paper takes regular tetrahedral particle groups as the research subject to analyze their motion and distribution law in liquid-solid two-phase flow in vertical uppipe. In the investigation, the effects of particle inlet volume fraction (1%, 2%, 3%, 4% and 5%) and liquid inlet velocity (1.0, 1.2, 1.5, 1.8 and 2.0 m/s) on the average velocity and relative volume fraction distribution of particle groups in tubes are simulated based on CFD-DEM (Computational Fluid Dynamics-Discrete Element Model) coupling method, and the accuracy of the numerical simulation is verified by PIV (Particle Image Veloci-metry) experiments. The results show that, within the studied parameter range, the average velocity of particle groups exhibits axial fluctuations, with a fluctuation amplitude intensifying as the liquid inlet velocity increases, and decreases radially from the pipe center to the wall. Furthermore, the velocity distribution becomes increasingly centralized as the fluid flow develops axially. Along the radial direction, the relative volume fraction of particles follows the double peak law, that is, being higher in the central area and near the wall of the tube, while being lower in the transition area. When the particle inlet volume fraction is 1% and the liquid inlet velocity is 2.0 m/s, the particle volume fraction near the pipe wall is the highest.

Fusion Transformer Model-Based Segmentation Algorithm for Laser Point Cloud of Distribution Lines



DAI Zhou, LIU Yan, MAO Xianyin, GUO Tao, XU Lianggang, CHENG Guixian

2025, 53(5):  139-146.  doi:10.12141/j.issn.1000-565X.240542

Asbtract ( 1568 )   HTML ( 7)   PDF (1833KB) ( 36 )  

Figures and Tables | References | Related Articles | Metrics

As laser point cloud models are crucial for distribution line inspection and management, most distribution channels have constructed laser point cloud models at present. With the increase of the number of models, extracting key component locations (e.g., conductors, insulators) becomes vital. In order to enhance the accuracy and efficiency of segmenting key components such as lines, towers and insulators, this paper presents a segmentation algorithm for laser point cloud of distribution lines based on a fusion Transformer model. Given the need for detailed features in the point clouds of distribution lines, a dual-channel parallel feature extraction module is designed to capture high-frequency and low-frequency features. The low-frequency features are processed via average pooling and a fusion Transformer-based extractor, while the high-frequency features are handled through max pooling and a multi-layer perceptron (MLP) module with convolutional layers. The feature vectors from both channels are then fused to improve the ability of detail feature extraction. Additionally, the fused features are fed back into the MLP module for further refinement, achieving precise point cloud target segmentation. Extensive experiments demonstrate the accuracy and effectiveness of the proposed algorithm. It has potential advantages in many aspects, such as improving the inspection accuracy of unmanned aerial vehicles, enhancing the level of automation, improving the robustness, integrating multi-source data and reducing inspection costs.