Loading...

Archive

    For Selected: Toggle Thumbnails
    A Multi-target Pedestrian Tracking Algorithm Based on Generated Adversarial Network
    WEI Ying, XU Chu-qiao, DIAO Zhao-fu, LI Bo-qun
    2020, 41 (12):  1673-1680.  DOI: 10.12068/j.issn.1005-3026.2020.12.001
    Abstract ( 950 )   HTML   PDF (1298KB) ( 1187 )  
    In the field of multi-target tracking, the problems of target loss, identity exchange and switch are easy to occur under the conditions of complex background, target occlusion, target scale and attitude change. To solve these problems, a multi-target tracking algorithm was proposed based on detection. A human body and face association algorithm based on YOLO was used to classify and detect the position of the current frame objects, and the feature extraction model based on generative adversarial network was proposed to learn the main features and subtle features of the objects. Then the generative adversarial network was used to generate the motion trajectories of multiple targets, and finally the target’s motion and appearance information were merged to obtain the optimal matching of the target tracking results. The experimental results show that the multi-target tracking algorithm proposed is both accurate and robust. Compared with the current algorithms with the least ID switch, the number of ID switch is 65% less and the accuracy is improved by 0.25%.
    References | Related Articles | Metrics
    Information & Control
    Research and Implementation of Speech Emotion Recognition Based on CGRU Model
    ZHENG Yan, CHEN Jia-nan, WU Fan, FU Bin
    2020, 41 (12):  1680-1685.  DOI: 10.12068/j.issn.1005-3026.2020.12.002
    Abstract ( 852 )   HTML   PDF (1007KB) ( 781 )  
    Speech emotion recognition is a very important research direction in emotion computing and human-computer interaction. At present, deep neural network is widely used to extract emotional features of speech, but further research is needed on which neural network model to use and how to alleviate the problem of model overfitting. To solve these problems, a CGRU model was proposed, which combined one dimensional convolutional neural networks (CNN) and gated circulation unit (GRU). The low-order and high-order emotional features of speech were extracted from the MFCC features of the original speech signal, and the features were selected through random forest, which achieved 79%, 69% and 75% recognition accuracy respectively on three common emotional corpus: EMODB, SAVEE, RAVDESS. By using the data augmentation technique, the sample size was increased by adding gaussian noise and changing the speed, which further improved the identification accuracy. The availability of the model in the real world was verified through the online identification system.
    References | Related Articles | Metrics
    Research on the Critical Parameters for Droplet Formation in Co-Flow Based on Level Set Method
    HU Sheng, LIU Xin-qi, LYU Xiao-yong, WU Dong-xu
    2020, 41 (12):  1686-1691.  DOI: 10.12068/j.issn.1005-3026.2020.12.003
    Abstract ( 576 )   HTML   PDF (668KB) ( 656 )  
    The generation of droplets in co-flow was studied by the level set method. Taking into account liquid-liquid two phases, the dispersed phased liquid is a mixture combined paraffin oil with heptane, whereas, the continuous phase fluid is aqueous water. Herein, these calculated works were carried out with inlet velocity and viscosity of the dispersed phase fluid, as well as injected hole size. When the dispersed phase fluid with velocity from 15mm/s to 48mm/s flows into a hole channel with radius of 1mm, the jetting height increases and the periodic generation reduces as the inlet velocity increases gradually. Meanwhile, increasing its viscous coefficient could decrease the disinfected time between droplet tension and drug force, so that frequency was increased and homogeneous characteristics were improved. The velocity of laminar flow was directly affected by the size of the hole. Owing to shrink the hole size, the sharpness of injected velocity occurs, resulting in the smaller size of the extruded droplet.
    References | Related Articles | Metrics
    Intensity Trend Forecasting of the ENPEMF Signal Before Earthquake Based on Chaotic Parameters Optimized RBF Algorithm
    HAO Guo-cheng, GUO Juan, TAN Song-yuan, ZENG Zuo-xun5
    2020, 41 (12):  1692-1698.  DOI: 10.12068/j.issn.1005-3026.2020.12.004
    Abstract ( 583 )   HTML   PDF (1258KB) ( 545 )  
    A chaotic parameter-optimized radial basis function (RBF) forecasting model was proposed. The chaos theory was used to obtain the embedded dimension and delay time of the ENPEMF, and the obtained parameters were used to optimize the RBF neural network. Finally, the trained optimized-RBF was utilized to forecast the strength trend of 14 d ENPEMF data. Numerical simulation results show that the improved RBF model could forecast the Rossler time series well with small error. Applying the improved RBF algorithm to the ENPEMF data before Ms7.0 earthquake in Lushan, it can effectively forecast the ENPEMF intensity trend 14 d before earthquake. The forecasting effect and accuracy are significantly better than that of the traditional RBF algorithm, which is expected to provide support for electromagnetic monitoring and analysis before earthquakes and geological disasters.
    References | Related Articles | Metrics
    Simulation-based Solution for Multi-crane Dynamic Scheduling in Steelmaking Shop
    LI Ji, XU An-jun
    2020, 41 (12):  1699-1707.  DOI: 10.12068/j.issn.1005-3026.2020.12.005
    Abstract ( 672 )   HTML   PDF (893KB) ( 926 )  
    A simulation-based solution under the rolling scheduling strategy was provided for a multi-crane scheduling problem deriving from the steelmaking shop. The objective function of the model is to minimize the temporal deviation of the crane scheduling from the process scheduling and the workloads of the cranes. The constraints feature the non-interference constraints between cranes, the station capacity constraint and the constraints caused by varied initial conditions. The model was solved by a simulation-based heuristic. The simulation experiments were conducted based on a twin-crane scheduling instance in a steelmaking shop. The results show that the crane scheduling solution can maintain the stability of the production schedule. In addition, it can also balance and reduce the workloads of the cranes. In the rolling scheduling strategy, the solution obtained by the proposed heuristic method can reach, at most, 96.3% of the optimal one in optimization performance. Compared with the real-time rule scheduling method employed in the production field, the optimization performance was improved by 26.4% with the proposed heuristic method under the rolling scheduling strategy.
    References | Related Articles | Metrics
    Materials & Metallurgy
    Microstructure and Mechanical Properties of the Stirred Zone of Ultrasonic Assisted Friction Stir Welded Joint of 7075-T6 Alloy
    ZHANG Zhi-qiang, HE Chang-shu, ZHAO Su, ZHAO Xiang
    2020, 41 (12):  1708-1714.  DOI: 10.12068/j.issn.1005-3026.2020.12.006
    Abstract ( 729 )   HTML   PDF (6911KB) ( 766 )  
    Insufficient metal flow in the stirred zone(SZ) may easily lead to loose defects in the transition zone between the shoulder-affected zone and the pin-affected zone, and deteriorate the mechanical properties of the joint. 7075-T6 aluminum alloy was subjected to conventional friction stir welding (FSW) and ultrasonic assisted friction stir welding (UAFSW) under different welding parameters, and the effects of ultrasonic vibration on the metal flow behavior, microstructure characteristics and mechanical properties of the SZ were investigated. The results show that the mechanical properties of the SZ of UAFSW joints are better than that of FSW joints processed using the same welding parameters. Tensile strength and elongation of the SZ of UAFSW joint with the welding parameters of 1000r/min-110mm/min reach the maximum values, i.e. 515 MPa and 17.3%, respectively. The application of axial ultrasonic vibration during FSW can significantly reduce yield stress and flow stress of the metal in the SZ, promoting the flow of plasticized metal along the normal direction of the plate. Eliminating the loose defects in the transition zone of the SZ and refining the microstructure are the main reasons for improving the mechanical properties of the joint.
    References | Related Articles | Metrics
    Numerical Study on the Ringly-Slotted Stub Hole Design for Prebaked Anode
    LI Tuo-fu, TAO Wen-ju, WANG Zhao-wen, KONG Ling-yu
    2020, 41 (12):  1715-1720.  DOI: 10.12068/j.issn.1005-3026.2020.12.007
    Abstract ( 730 )   HTML   PDF (1487KB) ( 801 )  
    For the industrial anodes in Hall-Héroult cells, electric current can not pass through the stub hole bottom because of the large initial cast iron-to-carbon air gap at the stub hole bottom. A ringly-slotted stub hole design for the anode was proposed to solve this problem. The effects of the proposed design on the anode and the corresponding mechanism were evaluated by numerical simulations. The initial cast iron-to-carbon air gap at the stub hole bottom in the industrial anode under the gravity was 1.2 mm, while that in the proposed design was reduced to 0.2mm. Consequently, the thermal expansion of the cast iron and the stub caused the air gap to close during operation, which resulted in the contact pressure at the bottom of stub hole, increased the electrical contact area and improved the current distribution. The ringly-slotted stub hole leads to an anode voltage reduction of 22 mV, but shows insignificant effect on the temperature distribution of anodes.
    References | Related Articles | Metrics
    Mechanical Engineering
    Preparation and Grinding Performance of High Magnetic Field Vitrified Bond CBN Grinding Wheel
    SHI Ying, WANG Xue-zhi, YU Tian-biao, WANG Wan-shan
    2020, 41 (12):  1721-1726.  DOI: 10.12068/j.issn.1005-3026.2020.12.008
    Abstract ( 744 )   HTML   PDF (1687KB) ( 707 )  
    In view of the random grain orientation and low strength of vitrified bond CBN wheel, a high magnetic field was introduced into the preparation process of grinding wheel. During the preparation process, the Ni-plating CBN abrasive grains were added. It was found that the appropriate magnetic field strength could achieve the deflection of the Ni-plating CBN. In addition, the appropriate magnetic field strength is conducive to improving the strength of vitrified CBN composite materials. When the magnetic field strength was 6 T, the bending strength of vitrified CBN composite materials was the highest, and the value reached 79.5 MPa. Finally, it is proved that the specific grinding energy of CBN grinding wheel with high magnetic field was slightly lower than that of ordinary vitrified bond grinding wheel when grinding titanium alloy TC4. This research may have important practical significance for improving the properties of CBN grinding wheel and exploring new preparation technology.
    References | Related Articles | Metrics
    Study on Surface Material Removal Uniformity in Double Side Grinding Based on Abrasive Particle Trajectories
    LI Qing-liang, XIU Shi-chao, ZHANG Peng, WANG Kun
    2020, 41 (12):  1727-1732.  DOI: 10.12068/j.issn.1005-3026.2020.12.009
    Abstract ( 659 )   HTML   PDF (879KB) ( 747 )  
    In the double side grinding process, the two surfaces of the workpiece were ground separately by two end grinding wheels, which has a high processing efficiency. However, the removal non-uniformity of surface materials still exists. Material removal distribution model was established and the simulation and experimental researches were carried out under different grinding parameters in order to analyze the factors affecting the removal non-uniformity of surface materials in the double side grinding. The research demonstrated that the speed ratio of the grinding wheel and the workpiece has a significant effect on the removal uniformity of surface materials. The simulation data was consistent with the experimental results, which provides a theoretical basis for improving the removal uniformity of surface materials and surface quality.
    References | Related Articles | Metrics
    Study on Vibration Characteristics of Mistuned Blisk Based on PIHISCMSM
    SUN Hong-yun, YUAN Hui-qun, ZHAO Tian-yu
    2020, 41 (12):  1733-1740.  DOI: 10.12068/j.issn.1005-3026.2020.12.010
    Abstract ( 621 )   HTML   PDF (1198KB) ( 641 )  
    In order to investigate coupling vibration characteristics of new structure aero-engine disc-blade-shaft integration system, an improved hybrid interface substructural component modal synthesis method with prestress (PIHISCMSM) was proposed, and the accuracy of dynamics analysis results was verified by modal analysis. Based on the Isight software integrated APDL language, a CAD/CAE parameter driven finite element analysis platform for blisk system was established. The dynamic characteristics of mistuned blisk system with different blade thickness was studied. The results show that, under the same mistuning pattern, for blisk system with different blade thickness, the sensitivity degree of modal localization factor to different modal order is different. With the increase of blade thickness, the frequency of blade dominant vibration increases, the resonance frequency of blisk system increases, the resonance peak decreases, and the increase of blade thickness can reduce the vibration response localization degree of the mistuned blisk system.
    References | Related Articles | Metrics
    Influence of Welding Sequence on Residual Stress and Deformation of T-Joint of Q690 Steel
    LAN Liang-yun, SHAO Guo-qing, ZHANG Yi-ting, WANG Yi-nan
    2020, 41 (12):  1741-1746.  DOI: 10.12068/j.issn.1005-3026.2020.12.011
    Abstract ( 680 )   HTML   PDF (1353KB) ( 712 )  
    The influence of welding sequence on residual stress and deformation of T-joint of Q690 steel was analyzed by using ABAQUS. The simulation models of temperature field and stress field of K-groove double-sided two-layer two pass welding was established. The calculation results show that the welding sequence has a certain influence on the residual stress and deformation. For the residual stress, the scheme 3, in which one side was welded first and then the other side, produced the minimum residual stress, and the high stress range of the weld area was also small. The overall stress level of alternative welding scheme 1 and scheme 2 was slightly larger, and the high stress range of the weld area was also larger. For welding residual deformation, the deformation of scheme 1 and scheme 2 was smaller, and the deformation of both sides was symmetrical, and the deformation of scheme 3 was the largest, accompanied by the overall tilt deformation.
    References | Related Articles | Metrics
    Analysis on Thermal Contact Resistance of the Connecting Structure in Enclosed Rugged Computer
    MA Guo-jun , MU Cui-cui, GUAN Xiao-le, WU Cheng-wei
    2020, 41 (12):  1747-1753.  DOI: 10.12068/j.issn.1005-3026.2020.12.012
    Abstract ( 706 )   HTML   PDF (1975KB) ( 493 )  
    The present work focuses on the thermal contact resistance (TCR) at the interface between the cooling board and the chassis rail in the enclosed rugged computer (ERC). Two types of connecting structures (3J and 5J structures) used in the ERC are studied using the finite element method (FEM) analysis software. Firstly, the contact pressure distribution and TCR distribution at the interface between the cooling board and the chassis rail are both studied. It is found that the TCR in 5J structure is much smaller than that in 3J structure. Secondly, the influences of bolt torque and surface roughness on the TCR of 5J structure are studied. The results show that TCR decreases with increasing bolt torque or decreasing surface roughness. However, a high bolt torque would result in a high contact pressure what may lower the service life of electronic devices. Further analysis shows that low stiffness of the chassis rail results in an inhomogeneous contact that leads to an inhomogeneous TCR distribution and high average TCR. Based on these findings, an optimization on the chassis rail is carried out to reduce both the average TCR and the contact pressure successfully. Moreover, the results show that the decrease in average TCR induced by the above optimization can be enhanced by increasing bolt torque or decreasing surface roughness.
    References | Related Articles | Metrics
    CFD-based Performance Analysis of a Discharge Valve Plate in the Rolling Rotor Compressor
    DENG Wen-juan, ZHANG Ying-li, DING Jia-nan, BA De-chun
    2020, 41 (12):  1754-1759.  DOI: 10.12068/j.issn.1005-3026.2020.12.013
    Abstract ( 677 )   HTML   PDF (1554KB) ( 897 )  
    Based on the CFD(computer fluid dynamics) model of the verified rolling rotor compressor, the periodic gas load of the valve plate is obtained, and the transient dynamic analysis and fatigue life prediction of the discharge valve plate are carried out by using the Workbench software. The influence of the geometric parameters of the valve on the flow area and the life of the valve plate is analyzed. It is found that the flow area and fatigue strength of the valve plate are related to the thickness of the valve plate and the characteristic lift of the valve plate. Under the specific conditions with the valve plate thickness of 0.305mm and the characteristic lift of the valve plate of 2.5mm, ideal performance results are obtained while the flow area and fatigue life of the valve plate are taken into account too. On this basis, the retainer profile is optimized, the flow area of the optimized valve plate increases by 8.63%, and the maximum stress decreases by 107.4MPa. The research results provide an effective reference and evaluation method for the design of discharge valve plate of rolling rotor compressor.
    References | Related Articles | Metrics
    Numerical Modeling and Analysis of Laser Polishing Based on Ablation Principle
    XU Ji-lin, ZOU Ping, WANG Wen-jie, YANG Xu-lei
    2020, 41 (12):  1760-1766.  DOI: 10.12068/j.issn.1005-3026.2020.12.014
    Abstract ( 776 )   HTML   PDF (1270KB) ( 788 )  
    A CW laser was used to perform laser polishing experiments on the 304 stainless steel based on the ablation principle, and a good polishing effect is obtained. A two-dimensional transient numerical model was developed by coupling the heat transfer and material vaporization in laser polishing. The evolution of the workpiece surface in laser polishing was simulated. The surface morphology and surface roughness of the workpiece after laser polishing were predicted. The simulation results are close to the experimental ones, and the error is only 8.17%. In addition, this paper studied the distribution of the surface moving velocity of the workpiece in laser polishing, and determined the relationship between the surface moving velocity and the initial surface morphology: the surface moving velocity at the peak position of the surface morphology is larger, while the surface moving velocity at the valley position is smaller, which reveals the mechanism of laser polishing based on ablation principle to reduce the surface roughness of the workpiece.
    References | Related Articles | Metrics
    Resources & Civil Engineering
    Analysis on Catastrophic Sliding of Pillar-Roof Supporting System in a Sedimentary Bauxite Goaf Mine
    JIANG Li-chun , XIE Bo
    2020, 41 (12):  1767-1774.  DOI: 10.12068/j.issn.1005-3026.2020.12.015
    Abstract ( 536 )   HTML   PDF (2480KB) ( 630 )  
    Based on the instability characteristic of a sedimentary bauxite goaf mine, which showing a shear failure along weak structural planes, a catastrophic sliding model of a pillar-roof supporting system was established by using the energy dissipation theory and the catastrophe theory. The instability mechanism related to the weakness structural plane was studied, and the influence of various internal control factors on the stability of the goaf was analyzed. The results indicate that the sudden sliding instability of the goaf is caused by the combined action of the internal and external control factors of the supporting system. When the external control factors are fixed, the sensitivity of each internal control factor can be listed from large to small as: goaf span b, pillar width a, thickness of weak zone B, structural weakness plane angle θ, pillar yield zone width c. The sensitivity of the thickness B of the weak zone and the inclination angle θ of the weak structural plane are just slightly smaller than that of the goaf span and the pillar width, which means that the influence of the weak structural plane should be paid more attention in actual projects. The reliability of the calculation results was verified in the Chenguang bauxite project. The research results can provide basis for stope parameter design of such mines.
    References | Related Articles | Metrics
    Component-based 3D Modeling Method for Complex Mine Roadway Network
    JIA Qing-ren, CHE De-fu, ZHONG Ren-qing
    2020, 41 (12):  1774-1780.  DOI: 10.12068/j.issn.1005-3026.2020.12.016
    Abstract ( 836 )   HTML   PDF (1435KB) ( 899 )  
    The modeling of complex mine roadways, such as curved roadways, intersection roadways, and variable cross-section roadways, is the key and difficult point when building the 3D models of roadway network. Deficiencies exists in the existing method when reconstructing complex roadway networks based on traverse data, especially for the curved intersection roadways which are common in mines. This paper proposed a basic component assembly method (BCAM) for the modeling of the 3D roadway network by using the underground survey data of coal mine roadway. By applying three steps of BCAM, namely the modeling of roadway centerlines, the modeling of roadway sidelines, and the modeling of the basic component of roadway (BCR), 3D models of complex roadway network can be automatically constructed. The experimental results show that the proposed method is robust, and the modeling time increases linearly and steadily with the increase of data volume. The established 3D model of the roadway network can support various analyses on space, geometry, and mechanics for smart mines.
    References | Related Articles | Metrics
    Management Science
    Method for Large Group Decision-Making with Multiple Elements Considering Multiple Subgroup Opinions
    LI Ming-yang, WANG Shuang, CAO Ping-ping, ZHAO Xiao-jie
    2020, 41 (12):  1781-1787.  DOI: 10.12068/j.issn.1005-3026.2020.12.017
    Abstract ( 621 )   HTML   PDF (409KB) ( 522 )  
    A decision analysis method is proposed for large group decision-making with multiple elements. Firstly, the borda score of each element for each participant is calculated according to the rankings of the constituent elements of the alternatives by the individuals participating in decision-making. Secondly, the average borda score vector of each subgroup is constructed and normalized as a virtual alternative for each subgroup. Then, the weight of each subgroup is calculated considering the differences in the consensus of evaluations of each subgroup. Furthermore, the similarity degree between each virtual solution and each alternative is calculated and aggregated with the weight of each subgroup to obtain the comprehensive similarity degree for each alternative, and then the alternative is ranked accordingly. Finally, the feasibility of the proposed method is illustrated by an example of urban park design alternative selection considering the public will.
    References | Related Articles | Metrics
    Value Evaluation of Innovative Technical Talents in Enterprises Based on Entropy Weight TOPSIS
    WANG Xin, CHEN Ni, ZHAO Ya-wen
    2020, 41 (12):  1788-1793.  DOI: 10.12068/j.issn.1005-3026.2020.12.018
    Abstract ( 560 )   HTML   PDF (371KB) ( 699 )  
    Using the iceberg model for reference to classify competence, the value of innovative technical talents is divided into explicit value and implicit value. The explicit value includes innovation knowledge, innovation skills and innovation achievements while the implicit value includes innovation character, innovation motivation and innovation potential. Based on the results of literature research and in-depth interviews, 21 evaluation indexes of innovative technical talents are selected, the weight of indexes is determined by the entropy weight method, and the value of innovative technical talents is evaluated by the TOPSIS method. The evaluation process is realized by MATLAB software, and the value evaluation system of innovative technical talents in enterprises is established. The results show that the TOPSIS method based on entropy weight can effectively evaluate the element value and comprehensive value of innovative technical talents in enterprises, and provide an important basis for the selection, training and stimulation of talents in enterprises.
    References | Related Articles | Metrics
    Influence of Stock Index Futures on Positive Feedback Trading in the Stock Market: An Econometric Test in China’s A-Share Market
    TIAN Shu-xi, HU Jing-xue, SUN Ying, WANG Jian
    2020, 41 (12):  1794-1799.  DOI: 10.12068/j.issn.1005-3026.2020.12.019
    Abstract ( 517 )   HTML   PDF (374KB) ( 515 )  
    Based on the asymmetric generalized autoregressive conditional heteroscedasticity (GARCH) model and positive feedback trading model, the positive feedback trading in China’s A-share market is tested. The results show that positive feedback trading in the stock market reduces after the launch of stock index futures, whose cause lies not in the improvement of market information efficiency, but in the positive feedback trading migration from the stock spot market to its futures market, that is to say, stock index futures play a role mainly in the transference of positive feedback trading. This conclusion is different from the previous optimistic evaluations on China’s stock index futures. The results also show that investors’ risk aversion has a negative correlation with the intensity of positive feedback trading, that is to say, positive feedback trading shows significant risk seeking preference. Although the strict constraints on stock index futures may reduce the coupling effects between the stock spot market and its futures market, it can’t change the risk seeking preference for positive feedback trading, so the relevant system arrangements should focus on the improvement of information efficiency instead of volatility suppression.
    References | Related Articles | Metrics
    Mathematics
    Parabolic Helicoidal Surfaces in Affine Spaces
    YU Yan-hua, PENG Lan-lan, JIA Kun
    2020, 41 (12):  1800-1804.  DOI: 10.12068/j.issn.1005-3026.2020.12.020
    Abstract ( 646 )   HTML   PDF (674KB) ( 676 )  
    Helicoidal surface is a surface formed by a plane curve that rotates around a fixed axis while uniformly shifting along the direction of the axis. Helicoidal surface can be divided into elliptic, hyperbolic and parabolic types based on different rotation axis. The helicoidal surface of parabolic type is studied by using Blaschke metric in affine spaces. The helicoidal surfaces of parabolic type whose Gaussian curvature and mean curvature vanish identically have been classified, respectively. When h=0, the helical motion degrades into rotational motion. In this situation, the flat and minimal parabolic surfaces of revolution are also classified. Finally, some graphs of those surfaces are given.
    References | Related Articles | Metrics