Information
Edited & Published: Editorial Office of Journal of Dalian Jiaotong University
Publication Period: Bimonthly
Publisher: Dalian, Liaoning Province, China
Language: Chinese
Superintendent: Liaoning Provincial Department of Education
Sponsored by: Dalian Jiaotong University
Editor in Chief: Chen Bingzhi
Executive Editor in Chief: Guo Ruijun
China National unified serial number: CN 21-1550/U
International Standard Serial Number: ISSN 1673-9590
Periodical website: https://dltd.cbpt.cnki.net
Related Links
Page Views
Highway Traffic Flow Prediction Based on a Feature Fused Spatio-Temporal Graph Mixed Networks
ZHANG Yang;ZHOU Chenfeng;CHEN Yanlin;Mining the spatio-temporal correlations between different nodes is the key to improving the accuracy of highway traffic prediction. However, spatial modeling is limited by methods that only consider static spatial dependencies or dynamic spatial correlations. To solve the above problems, a feature fusion spatio-temporal graph mixed network traffic flow prediction model(FF-STGM) is proposed. A multi-attention mechanism module is introduced, the spatio-temporal features of traffic flow and the structural attributes of the freeway network are structurally extracted using a doubly-parallel optimized spatio-temporal graph convolutional network, and the non-continuous time dynamic correlations in the data are captured. Meanwhile, an optimized adaptive ant colony optimization algorithm based on the improved Lévy flight strategy is used to adjust dynamically the structural parameters of the model to optimize the hidden layer structure of the spatio-temporal graph mixed network. The experiment results show that the MAE values are reduced by 6.45, 7.69, and 16.8 compared with STSGCN, ASTGCN and CNN-LSTM. The improved spatio-temporal graph mixed network model prediction accuracy outperforms other comparative models.
Elderly Public Transportation Mode Choice Model Considering Individual Heterogeneity
CAO Yi;ZHAO Bin;ZHOU Zeyu;To uncover the travel characteristics and patterns of the elderly and facilitate the age-friendly reform and development of urban public transportation systems, this study investigates the influencing factors and mechanisms underlying the elderly choice of public transportation modes. By integrating SP(Stated Preference) and RP(Revealed Preference) survey methods, panel data necessary for regression analysis were obtained to analyze the basic travel characteristics of the elderly. The survey revealsa slight male majority, with participants predominantly aged 60-69(64.59%), mostly retired workers(59.73%) in good health but with some illnesses. Most households have2-3 children(47.84%) and own a car(71.9%). Elderly travelers often travel in groups without children, tolerated waiting times of 15-30 minutes, walk for over 5 minutes, travel within 10 km, and preferred fewer transfers but in larger groups. Considering individual heterogeneity, the comfort, convenience, and reliability of transportation modes were used as psychological latent variables, and SEM(Structural Equation Modeling) was employed to calibrate latent variable parameters. Validity, reliability, fit indices, and path analysis confirmed the SEM model accuracy in representing the latent variables. The coefficients of travel cost and walking distance were determined as random parameters using stepwise regression, leading to the development of a SEM-Random Parameter Logit model. Compared to the models without SEM and traditional Logit ones, the SEM-Random Parameter Logit model hasthe lowest AIC and BIC, indicating optimal fit. The study findsthat travel fare and walking distance havesignificant negative impacts on travel utility, but individual heterogeneity leadsto varying effects. The coefficients are influenced by individual characteristics, travel attributes and psychological latent variables of travel perception. Notably, travel time hasthe most significant negative impact on travel value, while convenience hasa significant positive effect.
Pharmaceutical Cold Chain Distribution Path Optimization of Electric Vehicles Based on Charging and Swapping Mode
WU Huirong;XU Zexin;OUYANG Chao;To address the phenomenon where different factors influence the distribution costs of electric refrigerated trucks under two power replenishment modes, the following approach is adopted.Aiming to minimize distribution costs, the optimization models for route planning are constructed separately for the charging mode which considers battery degradation costs and time-of-use electricity pricing and the battery-swapping mode which accounts for leasing fees and swap service costs. Then, the traditional genetic algorithm is improved, a rational power replenishment station insertion strategy is designed to reduce the frequency of replenishment and total travel distance, and a simulated annealing mechanism is incorporated to prevent premature convergence of the solution. Finally, experiments are conducted to validate the algorithm feasibility, comparing distribution costs and time under different charging rates and replenishment limits. The results demonstrate that the battery-swapping mode yields lower distribution costs, and the lowest distribution cost is achieved when the replenishment limit is set at 80%.
Bearing Fault Identification Method Based on EEMD-Envelope Spectrum and JS-SDAE
YUAN Yu;GUO Qi;In order to solve the problem of difficulty in identifying multiple states of different damage positions and degrees of rolling bearings, a bearing fault diagnosis method based on Ensemble Empirical Mode Decomposition(EEMD) envelope spectrum and artificial Jellyfish Search(JS) optimizer-SDAE(Stacked Denoising Auto Encoder) was proposed. Firstly, EEMD was used to decompose the bearing signal to retain the intrinsic mode function(IMF) that correlates with the original signal height. Secondly, the envelope spectrum of the selected components is used to construct high-dimensional features as the input of the network. Finally, the dimensionality reduction was input into the SDAE optimized by the artificial jellyfish optimization algorithm to complete the multi-class fault identification of the bearing. Experiments show that after 10 types of feature data are input into SDAE for learning, the EEMD envelope spectrum can better reflect the fault characteristics than the time-domain signal. Moreover, the JS-SDAE network has higher accuracy than the decision tree, Bayesian, grid search optimized Bayesian, SVM, Bayesian optimized SVM, KNN, Bayesian optimized KNN and other algorithms. The data collected by the QPZZ-II system acquisition experiment platform were used for verification, and the results show that the model test set reaches an accuracy rate of 96.7%.
Analysis of Influencing Factors of Fatigue Failure Mode of Load-Bearing Fillet Weld
XIE Suming;WU Yingchao;CHENG Yajun;NIE Chunge;In order to study the fatigue failure conditions of fillet weld toe under tensile and bending loads, the factors affecting the structural stress of the weld throat and weld toe section of load-bearing fillet weld were analyzed by analytical method and structural stress method. Firstly, the analytical expressions of normal structural stress, tangential structural stress and equivalent structural stress of any weld throat section with or without penetration depth are derived based on the structural stress. The failure angle of the weld throat is obtained by striving for the extreme value of equivalent structure, which is increased with the increase of penetration depth. Secondly, the structural stress of the weld throat and weld toe section of the fillet weld of the cross joint subjected to pulsating cyclic tensile load is calculated. The results show that the structural stress of the weld throat section is greater than that of the weld toe section, and the fatigue life number of the weld throat failure is 1.21×105, which is consistent with the experiment results. Then, the failure analysis of the fillet weld of the cross joint finds that the weld throat fails when there is no penetration, and the weld toe fails when there is full penetration and is not affected by the size of the weld toe. When the size of the welding foot is constant, the welding throat fails when p/t<0.36, and the welding toe fails when p/t>0.36. Finally, the failure analysis of the single-sided fillet weld under bending load finds that when the weld angle size is constant, the weld throat failure occurs when p/t<0.17, and the weld toe failure occurs when p/t>0.17.
WeChat official
