Training Speed Research Articles

Bicycling for Rehabilitation of Persons With Parkinson Disease: A Scoping Review.

Exercise is beneficial for persons with Parkinson disease (PwPD). The overarching purpose of this scoping review was to provide guidance to clinicians and scientists regarding current evidence for bicycling exercise for PwPD. A scoping review was conducted to examine the heterogeneous literature on stationary bicycling for PwPD to reduce motor symptoms and body function structure impairments, improve activities and motor performance, and reduce disease severity. The Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines were followed. PubMed, CINAHL, and PEDro were searched from inception to January 23, 2023. Articles reporting original data on relevant outcome measures were included. Search results were screened and articles were extracted. Data were analyzed quantitatively with percentages of significant and clinically meaningful findings and qualitatively to extract themes. Bicycling was categorized using bicycle types (assisted, nonassisted) and training modes (speed, aerobic, force). A high percentage of the 34 studies showed statistical significance for reducing motor symptoms (83%), body function structure impairments (78%), disease severity (82%), and improving activities (gait 72%, balance 60%). Clinically meaningful findings were achieved in 71% of the studies for reduction in disease severity and in 50% for improving gait. The literature on bicycling for PwPD has evolved from speed to aerobic studies. The terminology describing types of bicycling was simplified. Of all the outcomes reported, reduction of disease severity achieved the highest frequency of clinical meaningful improvements. Bicycling was comparable with other forms of aerobic training for walking speed and endurance. Opportunities for translation to practice and research are presented. for more insights from the authors (see the Video, Supplemental Digital Content 1 available at: http://links.lww.com/JNPT/A462).

Theoretical study on the most unfavorable tunnel length (MUTL) based on the interior pressure variations of high-speed subway trains

When a high-speed subway passes through a tunnel, the external aerodynamic pressure of the train is transmitted into the train interior, which adversely affects the passenger comfort. The effect of tunnel length on the interior pressure variation of the train cannot be ignored. According to previous studies, there is a most unfavorable tunnel length (abbreviated as MUTL in this paper) for the exterior pressure of the train. However, MUTL for the interior pressure variation is still a research gap. In this study, the effect of tunnel length on the interior pressure variation of tunnel and train was investigated using numerical simulation and TWS method. Additionally, the theoretical formula of MUTL based on the internal pressure variation at the head and tail of the train was established and verified using the results of past field measurements and TWS method. The findings indicate that the interior pressure variation with tunnel lengths within a certain time interval is not monotonic. Moreover, there may be two or even more MUTLs, and the MUTL is also be influenced by the dynamic sealing index. MUTL decreases with increasing train speed and increases with increasing number of cars. The research findings of this paper address the research gap in the study of MUTL on the pressure variation inside trains. These findings offer new theoretical guidance for the research of aerodynamic effects in high-speed subway tunnels, and provide a valuable reference engineering design and optimization.

The effects of in-situ synthesized TiC on the performance improvement of nickel-based composite coatings for rail repair

With the increase in train speed and axle weight, the quality and service life of the rail (U71Mn steel) are more demanding. Aiming at the excellent properties of nickel/nickel-based alloys and ceramic materials, this paper prepares TiC/Ni45 composite coatings for rail repair by in-situ synthesis of ceramic phases using laser cladding technology. The effects of in-situ synthesized TiC on the physical phase, microstructure, microhardness and wear resistance were investigated and discussed in detail. The results show that the reaction of in-situ synthesized TiC can proceed positively, and the microstructure of the cladding layer mainly consists of the segregation of Cr, TiC particles, and substrate phase. With the increase of nickel-coated graphite and titanium powder, the average microhardness value of the cladding layer increased, and the average friction coefficient and wear volume decreased. When the mass fraction of (Ti + C) reaches 15 wt%, the average friction coefficient of the cladding layer reaches the minimum value. When the mass fraction reaches 20 wt%, the average microhardness of the cladding layer reaches 2.65 times that of the substrate, and the wear mechanism is transformed from mainly adhesive wear to abrasive, oxidative, and fatigue wear.

Study on the adaptive relationship between train nose lengths and variable-section tunnel parameters: Aiming at mitigating micro-pressure waves

Environmental problems caused by micro-pressure waves (MPWs) are increasingly prominent with increase in train speed. Considering different models of trains, which may have differences in nose lengths (Ln), and certain extreme terrains where it is unsuitable to construct the buffer hood outside the tunnel, this study numerically investigated the adaptive relationship between Ln and variable-section tunnel parameters [i.e., the length (L1) and area (S1) of the entrance expansion section] for minimizing the MPW. The results show that, as Ln rises from 5 to 13 m, the formation of the MPW peak is continuously delayed, and the peak value of the MPW decreases; while the optimal L1 remains at 20 m for different Ln. Additionally, the optimal S1 scarcely changes with the increase in Ln as well, between 151 and 156 m2, and its ratio to the section area of the main tunnel (70 m2) is approximately 2.2. The optimal parameters of the entrance expansion section play a universal role for trains with different Ln, which can alleviate the MPW amplitude with a value of 58% or higher. The findings can provide significant support for designing tunnels and tunnel hoods, focusing on alleviating acoustic noise.

Research on the drag reduction of high-speed train based on bottom two-multistage wing deflector

To address the challenge of reducing aerodynamic drag while further enhancing the speed of high-speed trains, this research employs the concept of flow control for the bottom parts and draws inspiration from the front wings of Formula 1 (F1) race cars. Three kinds of two-multistage wing deflectors are designed and systematically analyzed by unsteady Reynolds-averaged Navier–Stokes (URANS) turbulent model. The most suitable design is determined by the single bogie model with a simplified train body. Using the improved delayed detached eddy simulation method, the aerodynamic drag of 1:8 three-car train models with or without two-multistage wing deflector is studied at different operational speeds. The results present the total drag reduction is higher at higher speeds. The reductions of 4.26%, 3.92%, 3.63%, and 3.49% are obtained at the operating speeds of 400, 350, 300, and 200 km/h, respectively. The two-multistage wing deflector desirably improves the flow structure at the bottom of the train, which leads to the reduction of aerodynamic drag and a corresponding reduction in the positive pressure zones within the bogie area. Furthermore, the deflector restricts shedding vortices, effectively narrowing the interference range of airflow under the train, which will provide a potential drag reduction method for the next generation high-speed train.

FXAM: A unified and fast interpretable model for predictive analytics

Predictive analytics aims to build machine learning models to predict behavior patterns and use predictions to guide decision-making. Predictive analytics is human involved, thus the machine learning model is preferred to be interpretable. In literature, Generalized Additive Model (GAM) is a standard for interpretability. However, due to the one-to-many and many-to-one phenomena which appear commonly in real-world scenarios, existing GAMs have limitations to serve predictive analytics in terms of both accuracy and training efficiency. In this paper, we propose FXAM (Fast and eXplainable Additive Model), a unified and fast interpretable model for predictive analytics. FXAM extends GAM’s modeling capability with a unified additive model for numerical, categorical, and temporal features. FXAM conducts a novel training procedure called Three-Stage Iteration (TSI). TSI corresponds to learning over numerical, categorical, and temporal features respectively. Each stage learns a local optimum by fixing the parameters of other stages. We design joint learning over categorical features and partial learning over temporal features to achieve high accuracy and training efficiency. We prove that TSI is guaranteed to converge to the global optimum. We further propose a set of optimization techniques to speed up FXAM’s training algorithm to meet the needs of interactive analysis. Thorough evaluations conducted on diverse data sets verify that FXAM significantly outperforms existing GAMs in terms of training speed, and modeling categorical and temporal features. In terms of interpretability, we compare FXAM with the typical post-hoc approach XGBoost+SHAP on two real-world scenarios, which shows the superiority of FXAM’s inherent interpretability for predictive analytics.

The thermo-mechanical coupling method of brake disc based on dynamic convective heat transfer

With the increase of train speed, the brake disk receives more heat input, causing its own temperature to reach a higher level. Due to the complex structure of the brake disk, its convective heat transfer coefficient is difficult to calculate, which makes the analysis of the convective heat dissipation between the brake disk and the air complicated. In this article, based on the software FLUENT and ABAQUS, by combining the fluid-solid coupling method and the thermo-mechanical coupling method, the dynamic convective heat transfer and the transient temperature field of brake disk are simulated. The maximum value of dynamic convective heat transfer coefficient is 483 W/m2 ∙ K at the running speed of 350 km/h. As for the transient temperature field, not only the contact relationship between the brake disk and the brake pad is considered, but also the dynamic convective heat transfer between the brake disk and the surrounding air domain is applied. The simulation results are more in line with the actual situation and the engineering application.

GA-GGD: Improving semantic discriminability in graph contrastive learning via Generative Adversarial Network

Graph contrastive learning has garnered considerable research interest due to its ability to effectively embed graph data without manual labels. Among them, methods based on Deep Graph Infomax (DGI) have been widely studied and favored in the industry because of their fast training speed, applicability to large-scale data. DGI-based methods usually obtain a noise graph through node shuffling. The proxy task of these methods encourages the encoder to distinguish whether the nodes come from the original graph or the noise graph, thereby maximizing the mutual information between the node representation and the graph it belongs to, while also maximizing the Jenson-Shannon divergence between the nodes of original graph and noise graph. However, we argue that these approaches only enable the encoder to differentiate between semantically meaningful graphs and noise graphs, but not to effectively identify different semantic graphs. This leads to the inability of the encoder to effectively embed information between different semantics, significantly reducing the robustness and affecting the performance of downstream tasks. In addition, this training mode makes the model more sensitive to attacks. To improve their semantic discriminability, we take advantage of the natural ability of generative adversarial networks to generate semantic data, proposing a method called Generative Adversarial Graph Group Discrimination (GA-GGD). Specifically, it consists of a graph group discriminator and a semantic attack generator. The discriminator aims to encode the graph and identify whether nodes originate from the original graph. The goal of the generator is to use random features and graph structure to find vulnerabilities of discriminator and generate node representations with similar but wrong semantic to confuse the discriminator. GA-GGD can improve the model’s semantic information embedding without significantly increasing computational overhead and memory occupancy. We test the effectiveness of the proposed model on commonly used data sets and large-scale datasets, as well as in various downstream tasks such as classification, clustering, and adversarial attacks defence. A wealth of experimental results confirm the efficacy of the proposed model.

Devising a method for improving pedestrian traffic safety when crossing railroad tracks by implementing an information system with a fixed warning time

The object of research is pedestrian safety when crossing railroad tracks at pedestrian crossings. A method for improving the safety of pedestrian traffic when crossing railroad tracks within the area approaching the railroad pedestrian crossing has been developed, by introducing an information system for warning pedestrians about train movement with a fixed warning time. Experimental measurements of the speed of the rolling stock were carried out in the area approaching the pedestrian crossing over the railroad track. With the mixed movement of trains, it was found that the lowest speed of the train was 67.46 km/h, and the highest was the passenger train – 131.9 km/h. At the same time, the speed of freight trains ranged from 74 to 79 km/h. Such data are needed to determine the speed of the rolling stock on the approach to the crossing or transition, which will allow the light and sound signal to be turned on in time when the movement of a pedestrian across the track will be dangerous. The methodology for calculating the length of the approach section to the pedestrian crossing, depending on the speed of the rolling stock, has been given. It was established that the speed of the rolling stock has a significant effect on the length of the approach section. At the lowest measured speed of the train – 67.46 km/h, the length of the approach section was 317.46 m, and when the passenger train was moving at a speed of 131.9 km/h – 620.68 m. For a single-track section of the railroad, a fixed time has been established for warning pedestrians about the approach of a train to the crossing. It is a constant value that is included in the warning information system on a single-track section, and after it, a complex light and sound alarm with an information board is activated, which will warn the pedestrian about the movement of the train. The method for improving the safety of pedestrian traffic with a fixed time of the pedestrian’s anticipation of the train movement could be used at crossings with pedestrian traffic, as a supplement to the crossing signaling subsystem in railroad transport. It can also be used at equipped pedestrian crossings over railroad tracks in urbanized areas. The advantage of introducing such a method is a fixed time to warn the pedestrian, in each specific case of the train approaching the crossing

Scientific calculation and application of safety marking principles

In today's era, trains have become an indispensable mode of transportation for people, yet safety hazards are inevitable during the process of trains entering stations. To study these safety hazards and the principles behind the design of safety markings, we combine the Bernoulli Effect and the conservation of mechanical energy in fluids, taking into account factors such as a person's weight, volume, and distance from the train. We derive an expression that relates the pressure experienced by individuals to the train's speed, the train's cross-sectional area, the distance between the individual and the train, and the area of the individual subjected to force. By comparing this with the general friction force that corresponds to real-life situations, we find that the distances for safety markings provided by the national "Railway Technical Management Regulations" for high-speed train and dynamic train stations are reasonable: for speeds not exceeding 120km/h, the distance is 1000mm; for speeds over 120km/h but not exceeding 160km/h, the distance is 1500mm; and for speeds over 160km/h up to 200km/h, the distance is 2000mm, regardless of the platform height. Based on the factors affecting the safety markings on railway platforms, four practical safety recommendations are proposed.

Benchmarking core temperature forecasting for lithium-ion battery using typical recurrent neural networks

One of the factors hindering the scaled application of lithium-ion batteries is their thermal safety issues. Real-time monitoring of core temperature holds promise in alleviating this concern. However, calibrating individual models for thousands of battery cells poses challenges and leads to big data issues. The maturing and rapidly advancing artificial intelligence technologies hold promise in addressing the task of core temperature prediction. In this study, we conduct a benchmark test using typical recurrent neural networks for core temperature prediction. The test encompasses three main aspects: data acquisition, data processing, and data analysis. Firstly, we conduct dynamic battery tests across a wide temperature range on a custom-built experimental platform. Subsequently, we employ big data techniques for data pre-processing, including filtering, normalization, and sliding window methods. Next, hyperparameter optimization is then performed using the Bayesian optimization with Tree-structured Parzen Estimator, followed by K-fold cross validation. Consequently, we obtain three optimal performing models based on typical recurrent neural networks. Finally, we extensively compare these models in terms of performance and complexity. According to the results, all data-driven methods achieve satisfactory model performance, with RMSE within 0.17 °C and MAE within 0.15 °C. Among the models with similar predictive performance conditions, gated recurrent unit exhibits the lowest model complexity and faster training speed. Meanwhile, we also validate the portability and generalizability of the forecasting models. This study lays the foundation for future research in this area and can provide guidance for engineering practical applications to ensure the safe operation of battery energy storage systems.

Comparative study on lateral vibration characteristics of the head and tail cars of high-speed trains under unsteady aerodynamic loadings

With the increasing speed of high-speed trains, abnormal vibrations in the tail car pose a threat to its overall stability and running safety. In this study, the aerodynamic and multibody dynamics models for 3-car and 8-car trains are established. The flow field characteristics and dynamic performance around the head and tail cars under both steady and unsteady aerodynamic loadings at a speed of 400 km/h are compared. The results reveal that the surrounding flow field exhibits highly pronounced unsteady characteristics, directly leading to a significant impact of unsteady aerodynamic loadings on the train's dynamic response. Under unsteady loadings, the lateral stability indexes, displacement, velocity, and acceleration of the tail car are all worser compared to those of the head car. Additionally, the larger oscillation amplitudes and frequencies closer to the suspension modal frequencies under unsteady aerodynamic loadings contribute to more pronounced lateral displacements and accelerations, constituting a crucial factor in the observed ‘tail swing' phenomenon for the high-speed train.

A Static Security Region Analysis of New Power Systems Based on Improved Stochastic–Batch Gradient Pile Descent

The uncertainty in the new power system has increased, leading to limitations in traditional stability analysis methods. Therefore, considering the perspective of the three-dimensional static security region (SSR), we propose a novel approach for system static stability analysis. To address the slow training speed of traditional deep learning algorithms using batch gradient descent (BGD), we introduce an improved stochastic–batch gradient descent (S-BGD) search method that combines the advantages of stochastic gradient descent (SGD) in fast training. This method ensures both speed and precision in parameter training. Moreover, to tackle the problem of getting trapped in local optima and saddle points during parameter training, we draw inspiration from kinematic theory and propose a gradient pile (GP) training method. By utilizing accumulated gradients as parameter corrections, this method effectively avoids getting stuck in local optima and saddle points, thereby enhancing precision. Finally, we apply the proposed methods to construct the static security region for the IEEE-118 new power system using its data as samples, demonstrating the effectiveness of our approach.

Transfer‐Learning‐Enabled 3D reconfigurable broadband solar metamaterial absorbers design

Research on metamaterials shows excellent potential in the field of solar energy harvesting. In recent years, the design of broadband solar metamaterial absorbers (SMAs) has attracted significant interest with the wide application of deep learning methods. This paper proposes a deep neural network (DNN) to realize forward prediction and inverse design of reconfigurable 3D SMAs. In the inverse design, a polarization-insensitive broadband SMA with an absorption bandwidth of 2.7 μm and an average absorption rate of 97.6% with an adjustable bandwidth range of 369 nm is successfully designed. The design of SMAs with different structures is also realized by a transfer learning method to improve the training speed and performance further. Using the transfer learning approach, the training speed of the neural network target model can be accelerated, and its training performance can be improved on small datasets by utilizing the trained neural network source model. Meanwhile, using the trained inverse design target model, a polarization-insensitive broadband SMA was designed with an absorption bandwidth of 2.7 μm, an average absorption of 97.9%, and an adjustable bandwidth range of 141 nm. Finally, we verified the solar energy harvesting capability of the designed broadband SMAs under real-world conditions using air mass (AM) 1.5, and they were calculated to be capable of harvesting the vast majority of the energy. The method is instructive in the design process of SMAs and can be effectively used to explore multifunctional complex nanophotonic devices.

Design and optimization of haze prediction model based on particle swarm optimization algorithm and graphics processor

With the rapid expansion of industrialization and urbanization, fine Particulate Matter (PM2.5) pollution has escalated into a major global environmental crisis. This pollution severely affects human health and ecosystem stability. Accurately predicting PM2.5 levels is essential. However, air quality forecasting currently faces challenges in processing vast data and enhancing model accuracy. Deep learning models are widely applied for their superior learning and fitting abilities in haze prediction. Yet, they are limited by optimization challenges, long training periods, high data quality needs, and a tendency towards overfitting. Furthermore, the complex internal structures and mechanisms of these models complicate the understanding of haze formation. In contrast, traditional Support Vector Regression (SVR) methods perform well with complex non-linear data but struggle with increased data volumes. To address this, we developed CUDA-based code to optimize SVR algorithm efficiency. We also combined SVR with Genetic Algorithms (GA), Sparrow Search Algorithm (SSA), and Particle Swarm Optimization (PSO) to identify the optimal haze prediction model. Our results demonstrate that the model combining intelligent algorithms with Central Processing Unit-raphics Processing Unit (CPU-GPU) heterogeneous parallel computing significantly outpaces the PSO-SVR model in training speed. It achieves a computation time that is 6.21–35.34 times faster. Compared to other models, the Particle Swarm Optimization-Central Processing Unit-Graphics Processing Unit-Support Vector Regression (PSO-CPU-GPU-SVR) model stands out in haze prediction, offering substantial speed improvements and enhanced stability and reliability while maintaining high accuracy. This breakthrough not only advances the efficiency and accuracy of haze prediction but also provides valuable insights for real-time air quality monitoring and decision-making.

Self-sensing cementitious geocomposites in rail track substructures

A self-sensing cementitious geocomposite was developed based on laboratory data, and it demonstrated good physical and mechanical properties, durability, and piezoresistivity performance. It consisted of stabilized cemented sand containing multiwalled carbon nanotubes (MWCNTs) with graphene nanoplatelets (GNPs) as conductive fillers. This geocomposite could be used to detect damage based on the relationship between electrical impedance and mechanical performance, making it suitable for use as structural layers in railway lines. In this study, the effects of MWCNTs and GNPs, as well as of degree of saturation, were evaluated on the compaction, secant modulus, resilient modulus, electrical resistance, and piezoresistivity of the geocomposite. Scanning electron microscopy (SEM) and microscopic imaging were used to analyse microstructural alterations induced by varying concentrations of MWCNTs and GNPs. This innovative geocomposite, intended for installation in Portuguese railway lines as part of the EU project IN2TRACK3, is aimed at capturing performance data, identifying structural damage levels, and estimating load intensity, axle numbers, and train speed. The feasibility of its use is discussed based on literature studies and research conducted under the IN2TRACK2 and IN2TRACK3 projects, and its potential advantages over traditional methods for monitoring rail track health are highlighted.

Multi-objective optimisation of wheel-rail adhesion for high-speed trains with aerodynamic lift wings

A high-speed train with aerodynamic lift wings (aero-wings) is designed to improve train speed and mitigate energy consumption. Nevertheless, the impact of aero-wings on wheel-rail adhesion and safety remains uncertain. In this examination, a multi-body dynamics (MBD) model of a high-speed train integrated with aero-wings was first created. The influences of train speed, wheel-rail contact conditions, and aerodynamic lift on the wheel-rail adhesion, wheel wear, and dynamics performance were briefly investigated, indicating that aerodynamic lift degrades wheel-rail adhesion and dynamics performance but reduced wheel wear. Then, the multi-parameters matching optimisation design for sufficient wheel-rail adhesion and minimal wheel wear was conducted. The results indicate that, under dry contact and traction conditions, the minimum wear and optimal wheel-rail adhesion can be achieved when the train speed is about 440.62 km/h, the aerodynamic lift is about 24.48% of the vehicle weight, and the wheel wear optimisation rate is 25.38%. While the train speed of 549.54 km/h and an aerodynamic lift of 12.39% corresponds to the optimal results with wear reduction of 11.62%, under wet and braking conditions. This study provides valuable insights for determining aerodynamic load limits and formulating traction/braking control strategies for high-speed trains with aero-wings.

Redesign roadmap to support driving awareness at level railway crossings

The number of accidents at level railway crossings, especially crossings without gate barriers/attendants, is still very high due to technical problems, driving culture, and human error. The aim of this research is to provide road maps application based on ergonomic visual displays design that can increase awareness level for drivers before crossing railway crossings. The double awareness driving (DAD) map information system was built based on the waterfall method, which has 4 steps: defining requirements, system and software design, unit testing, and implementation. User needs to include origin-destination location, geolocation, distance & travel time, directions, crossing information, and crossing notifications. The DAD map application was tested using a usability test to determine the ease of using the application used the System Usability Scale (SUS) questionnaire and an Electroencephalogram (EEG) test to determine the increase in concentration in drivers before and immediately crossing a railway crossing. Periodically, the application provides information on the driving zone being passed; green zone for driving distances > 500 m to the crossing, the yellow zone for distances 500m to 100m, and the red zone for distances < 100 m. The DAD map also provides information on the position and speed of the nearest train that will cross the railway crossing. The usability test for 10 respondents giving SUS score = 97.5 (satisfaction category) with a time-based efficiency value = 0.29 goals/s, error rate = 0%, and a success rate of 93.33%. The cognitive ergonomic testing via Electroencephalogram (EEG) produced a focus level of 21.66%. Based on the results of DAD map testing can be implemented to improve the safety of level railroad crossings in an effort to reduce the number of driving accidents.

Triple Attention Mechanism with YOLOv5s for Fish Detection

Traditional fish farming methods suffer from backward production, low efficiency, low yield, and environmental pollution. As a result of thorough research using deep learning technology, the industrial aquaculture model has experienced gradual maturation. A variety of complex factors makes it difficult to extract effective features, which results in less-than-good model performance. This paper proposes a fish detection method that combines a triple attention mechanism with a You Only Look Once (TAM-YOLO)model. In order to enhance the speed of model training, the process of data encapsulation incorporates positive sample matching. An exponential moving average (EMA) is incorporated into the training process to make the model more robust, and coordinate attention (CA) and a convolutional block attention module are integrated into the YOLOv5s backbone to enhance the feature extraction of channels and spatial locations. The extracted feature maps are input to the PANet path aggregation network, and the underlying information is stacked with the feature maps. The method improves the detection accuracy of underwater blurred and distorted fish images. Experimental results show that the proposed TAM-YOLO model outperforms YOLOv3, YOLOv4, YOLOv5s, YOLOv5m, and SSD, with a mAP value of 95.88%, thus providing a new strategy for fish detection.

Research on vibration characteristics of a straddle-type station induced by moving trains

The novelty of straddle-type railway stations as a modern railway infrastructure necessitates thoroughly examining their vibration behavior under high-speed train-induced vibrations. This study focuses on Shangrao station, the inaugural straddle-type high-speed railway station in China, to investigate its vibration properties induced by high-speed trains. A comprehensive finite element model is developed and validated against empirical data to ensure accuracy. The vibration response and propagation patterns induced by railway operations are extensively studied under diverse operational conditions. The results illustrate that the peak acceleration in the vertical (Z) direction significantly exceeds that in the horizontal (X and Y) directions in most test cases. Moreover, under consistent train speed and axle load, the multiple parallel traffic scenario results in higher peak accelerations in all directions compared to the grade separation traffic scenario. Additionally, a positive correlation exists between the railway platform’s maximum amplitude, corresponding frequency, and train speed. Furthermore, increased train axle load leads to higher peak accelerations in the station building. The findings of this research offer valuable insights for the predictive modeling and design of high-speed straddle-type railway stations, providing essential technical support for future station projects.