Journal of Science and Transport Technology

Cross-Branch CNN-MLP Integration for Improving Landslide Spatial Probability on Mt. Umyeon, Korea

2025-11-06T03:37:29+00:00

Accurate maps of landslide spatial probability (LSP) are crucial for planning and risk reduction in steep, urbanized areas. A hybrid Convolutional Neural Network (CNN) - Multilayer Perceptron (MLP) model is introduced for mapping LSP in Mt. Umyeon, Korea. The design combines two complementary views of each location: a convolutional branch learns spatial context from multi-channel image patches. At the same time, a multilayer perceptron captures local numeric and categorical attributes at the central point. Feature-level fusion concatenates the embeddings from both branches and feeds a lightweight classifier to produce probabilities. Performance was assessed under consistent data splits and training protocols. Training AUCs reached 0.887 (MLP), 0.894 (CNN), and 0.903 (CNN-MLP). More importantly, validation AUCs were 0.808 (MLP), 0.821 (CNN), and 0.854 (CNN-MLP), indicating stronger generalization for the fused representation. These gains reflect the complementary nature of neighborhood structure learned by the CNN and pointwise information stabilized by the MLP. The results show that a compact, feature-level fusion of CNN and MLP can materially improve spatial probability mapping of landslides. The approach provides a practical route to more reliable probability surfaces for decision support in mountainous urban regions.

Free vibration of cracked multi-span continuous functionally graded nanobeams

2026-02-11T11:00:12+00:00

The free vibration of multiple cracked multi-span continuous functionally graded (FG) nanobeams is investigated based on Eringen’s nonlocal elastic theory (ENET). The massless three-spring model is employed to model transverse edged cracks. The governing equations for multiple cracked FG Euler–Bernoulli nanobeams are established by employing Hamilton’s principle, the ENET, and conditions of continuity at the crack locations. Analytical solutions are obtained to construct the dynamic stiffness matrix of nanobeam element. The proposed matrix enables an efficient and highly accurate free vibration analysis of cracked FG multi-span nanobeams while requiring only a minimal number of elements. The accuracy of the present approach is verified through numerical validations with previous works. The influences of nonlocal, material, and geometric parameters on the free vibration of multiple-cracked multi-span FG nanobeams are investigated in detail.

Interpretable Machine Learning Model for Evaluating Flexural Strength of Ultra High-Performance Concrete

2025-12-14T18:47:21+00:00

Ultra-high-performance concrete (UHPC) mix design remains experimentally expensive because many ingredients interact nonlinearly to govern flexural behavior. An interpretable machine-learning pipeline was developed to predict UHPC flexural strength from literature-derived mixes (317 observations, 14 input variables). Nine regression models were screened under a rigorous Monte-Carlo protocol (1,000 random 70/30 splits). Tree-based boosting dominated: on a representative split, CatBoost achieved R²test=0.928, RMSE = 1.980 MPa, MAE = 1.454 MPa, MAPE = 8.386%, with XGBoost close behind; Random Forest and Gradient Boosting formed a reliable second tier, while linear, SVR, and KNN underfit. Global and local interpretability (SHAP and PDP) revealed a stable hierarchy of drivers: steel fiber content and curing time were strongly beneficial; coarse aggregate content was deleterious and nearly monotonic; water became increasingly harmful at high dosages; superplasticizer exhibited an interior “sweet spot”; cement and silica fume were favorable (silica fume above ~100 kg/m³); sand was weakly positive; limestone powder was near-neutral. Guided by mean|SHAP|, the feature set was reduced from 14 to 9 variables with only a modest trade-off (R²test =0.916, RMSE = 2.143 MPa, MAE = 1.522 MPa, MAPE = 9.07%). External verification on an independent dataset confirmed generalization and preserved the correct nonlinear response to steel fibers in the practical 0-2% range. A lightweight GUI operationalizes the nine-input model, enabling rapid “what-if” exploration and reducing measurement burden by 36%. The results deliver both accuracy and transparency, distilling actionable rules for UHPC tailored to flexure-critical applications: prioritize steel fibers and adequate curing, cap coarse aggregate, and maintain water/superplasticizer within stable windows while using cement and silica fume to tune the matrix.

AI-Integrated BIM Education: A Conceptual Framework for Process Competencies Aligned with Industry Workforce Demands

2025-12-18T02:57:04+00:00

Building Information Modelling (BIM) has evolved from a visualisation aid to a process-driven methodology that demands interdisciplinary collaboration and rigorous information management aligned with ISO 19650. Yet many curricula still prioritise software proficiency over process understanding, leaving graduates under-prepared for BIM coordination, information management and decision-making. At the same time, rapid adoption of artificial intelligence (AI) in construction is reshaping BIM workflows and amplifying existing skills gaps. This conceptual paper develops a theoretically grounded framework for integrating AI into BIM education to cultivate both technical and process-oriented competencies. Drawing on the Technology Acceptance Model, constructivist learning theory and cognitive load theory, the framework positions AI as a cognitive scaffold that shifts students’ effort from routine modelling operations towards higher-order process reasoning. It specifies a scaffolded progression of AI use, authentic ISO 19650-aligned project work, collaborative interdisciplinary learning structures and assessment strategies that foreground process competencies rather than isolated software skills. The framework’s distinctive contribution lies in its explicit integration of three theoretical lenses, systematic mapping of learning outcomes to ISO 19650 and buildingSMART certification domains, and operational guidance through worked examples of AI-integrated instruction. Although conceptual and awaiting empirical validation, the framework offers actionable guidance for programme leaders and educators designing AI-enabled BIM curricula. It contributes to educational technology scholarship by illustrating how established learning theories can structure AI integration in technical education and by proposing AI as a pedagogical tool for addressing critical workforce development challenges in the construction industry.

GIS-Based Flow-R Model for Debris Flow Susceptibility Mapping: A Case Study from Muong Bo, Lao Cai, Vietnam

2025-12-16T01:28:42+00:00

Landslides and debris flows are frequent hazards in Vietnam’s mountainous regions, causing severe socio-economic damage that has intensified under climate change and extreme rainfall conditions. This study applies the open-source GIS-based Flow-R model to assess debris-flow susceptibility and runout characteristics in Muong Bo Commune, Lao Cai Province, northwestern Vietnam (specifically focusing on the highly vulnerable area around the Nam Cang stream). The methodology involved developing a geospatial database and simulating potential initiation and propagation zones. A major debris-flow event that occurred on 12 September 2023 was used for model calibration and validation. Field validation confirmed the model’s robust capability to simulate flow runout distances based on observed data. Seventy percent of the mapped debris-flow initiation points were used for calibration, and 30% for validation[/RED], and model performance was evaluated using the Area Under the ROC Curve (AUC). The Flow-R model achieved an AUC value of 0.868, indicating good predictive capability[/RED], while 48.5% of observed debris-flow initiation points were correctly predicted. Results demonstrate that Flow-R effectively delineates high-susceptibility source zones and plausible debris-flow runout paths, particularly for medium- to large-scale events. The novelty of this study lies in the first integrated application of Flow-R, combined with systematic field validation in northwestern Vietnam, and the coupling of detailed geological–geomorphological characterization with susceptibility and runout modeling, providing a transferable framework for debris-flow assessment in data-scarce mountainous regions.

A semi-analytical approach for nonlinear vibration of FG-CNTRC doubly curved shallow shells stiffened by FG-CNTRC stiffeners in thermal environment

2025-12-10T19:16:09+00:00

This paper investigates the geometrically nonlinear free and forced vibration behavior of functionally graded carbon nanotube-reinforced composite (FG-CNTRC) doubly curved shallow shells stiffened by FG-CNTRC stiffeners under harmonic pressure loads and in uniform thermal environment. The formulation is developed based on the higher-order shear deformation theory (HSDT) combined with von Kármán-type nonlinear kinematics. An enhanced smeared stiffener approach is employed to incorporate the contribution of stiffeners into the shell's global stiffness. The governing equations are derived using the Lagrangian method, with the Rayleigh dissipation function accounting for energy loss. The harmonic balance method is adopted to determine the stress function, and the nonlinear equations of motion are solved numerically using the Runge–Kutta method. Parametric studies are conducted to assess the effects of CNT distribution, curvature, stiffener orientation, and thermal loading. The results provide key insights into the nonlinear dynamic response of stiffened FG-CNTRC shallow shells, serving as a useful reference for design applications in thermomechanical environments.

Cementitious material based on Portland cement and ground granulated blast furnace slag for ground improvement of construction sites in the Mekong Delta

2025-11-19T08:42:45+00:00

This paper presents the results of a study on the selection of binder compositions based on Portland cement combined with ground granulated blast furnace slag (GGBFS) and gypsum for ground improvement using the deep mixing method (cement deep mixing – CDM) and shallow stabilization for road base and foundation works. The study focuses on weak and aggressive foundation soils in the Mekong Delta region. The binder compositions were determined through experimental investigations using three representative types of foundation soils (clayey sand, sandy clay, and high-plasticity clay) collected from expressway construction projects in the Mekong Delta region. Unconfined compressive strength (UCS) was evaluated at curing ages of 3, 7, 28, and 91 days, while durability performance was assessed through immersion in natural seawater up to 6 months. The results show that, relative to soils stabilized with ordinary Portland cement (PC40), the selected binder system—comprising PC40 combined with 55–65% GGBFS and 3–5% gypsum—resulted in a 28-day UCS increase of approximately 1.6 to 2.3 times. The stabilized soils with this binder also exhibited significantly improved durability when immersed in seawater. No cracking or failure was observed after 6 months of immersion, whereas the samples stabilized with ordinary Portland cement showed initial cracking after 3 months and complete failure after 6 months.

An analytical approach for nonlinear thermal buckling and postbuckling behavior of functionally graded graphene platelet-reinforced composite conical shells

2025-12-10T19:43:01+00:00

This study presents an analytical investigation into the nonlinear thermal buckling behavior of functionally graded graphene platelet-reinforced composite (FG-GPLRC) conical shells resting on a nonlinear elastic foundation. The formulation is developed based on Donnell shell theory in conjunction with von Kármán geometric nonlinearity. The nonlinear foundation is characterized by three stiffness parameters that capture both hardening and softening behaviors, corresponding to positive and negative nonlinear parameters, respectively. Employing the Ritz energy method, thermal load–deflection relationships are derived to analyze both the critical buckling temperature and the postbuckling response of the structure. The influence of key parameters, including the stiffness of the elastic medium, graphene platelet (GPL) mass fraction, material gradation profiles, and geometric configurations, on the nonlinear thermal buckling performance is thoroughly examined through numerical simulations.

Investigation of the control system for air supply during transient operation of a turbocharged diesel engine

2025-12-13T07:22:15+00:00

In the context of increasingly stringent emission standards and higher fuel efficiency requirements for turbocharged diesel engines, optimizing air-path control, particularly during transient operating conditions, has become essential. This study focuses on the design and evaluation of a dedicated air supply controller for transient operation to enhance the response and stability of a turbocharged diesel engine. The controller was developed to regulate the intake airflow dynamically through an auxiliary bypass line, compensating for turbocharger lag and maintaining stable boost pressure during rapid load or speed changes. Experimental results demonstrate that the proposed air supply controller significantly improves transient performance. Specifically, the transition time was reduced from 3 seconds to 1.5¸2 seconds, corresponding to an improvement of 16.6¸40%, while speed fluctuation Δn decreased by 8.5¸24%, depending on the load level. The best improvement was achieved at medium to high loads (50¸65%), indicating rapid and stable turbocharger response with the assistance of the supplemental air supply system. At low loads, the controller’s impact was less pronounced due to lower boost pressure and airflow rates. The results confirm that the proposed air-path controller for transient operation is a feasible and effective solution to shorten transient response time, reduce speed fluctuation, improve engine stability, and potentially reduce emissions in turbocharged diesel engines operating under rapidly varying conditions.

Free vibration and buckling analysis of bidirectional functionally graded plates with integrated piezoelectric layers

2025-11-23T09:02:35+00:00

In this paper, the vibration and buckling of functionally graded material plates with integrated piezoelectric layers are investigated. The material properties of the plate are assumed to vary along both the length and thickness directions (2D-FGM) according to a power-law while the electric potential in each piezoelectric layer varies linearly along the thickness direction. To perform this analysis, a finite element model based on the higher-order shear deformation refined plate theory (HSDT-4) and four-node rectangular element is developed. The governing equations are obtained by applying Hamilton's principle. The numerical results have demonstrated the convergence, accuracy and reliability of the established model. In addition, the influence of material parameters, geometry and mechanical boundary conditions on the vibration frequency and critical force of the 2D-FGM/PIE plate under both closed-circuit (Clocc) and open-circuit (Opcc) conditions is investigated and evaluated in detail.

GIS Based Soil Erosion Susceptibility Assessment Using Deep Learning Models: A Case Study in the Mountainous Region of Nghe An, Vietnam

2025-12-18T03:26:48+00:00

In this study, the main objective is to evaluate soil erosion susceptibility in the mountainous region of Nghe An Province using three deep learning models: Long Short-Term Memory (LSTM), Deep Neural Network (DNN), and Deep Attention Network (DaNet). A total of 685 erosion points were identified from field surveys and satellite imagery, and nine spatial conditioning factors were used, including slope, aspect, curvature, elevation, rainfall, Normalized Difference Vegetation Index (NDVI), soil type, distance to faults, and geology. The dataset was split into 70% for training and 30% for validating. Various validation metrics including area under the ROC curve (AUC) were used for validation and comparison of the models. The results show that among the tested models, DaNet showed the highest predictive performance, achieving an AUC of 0.936 on the training dataset and 0.852 on the validating dataset compared with other deep learning models (DNN and LSTM). The susceptibility map produced by DaNet demonstrated strong spatial alignment with real-world erosion occurrences, with 59.88% of observed erosion points located in the very high susceptibility class and 18.52% in the high class, totaling 78.4% of all erosion events. These results confirm DaNet’s effectiveness in capturing complex spatial patterns and delivering reliable erosion risk predictions, supporting its use for land-use planning.

Experimental evaluation of modulus of elasticity of concrete using limestone and basalt aggregates determined according to different standard test methods

2025-12-18T03:40:15+00:00

The modulus of elasticity of concrete is influenced not only by its compressive strength but also by the geological origin of the coarse aggregate used. However, current standards employ diverse testing methods and predictive equations, many of which neglect the aggregate’s contribution. This study presents experimental data for concrete with compressive strengths ranging from 30 MPa to 55 MPa, produced using two Vietnamese coarse aggregates: Ha Nam limestone and Hoa Binh basalt. The modulus of elasticity was measured according to three standards: EN 12390-13:2021, ASTM C469-10, and TCVN 5726-2022. At a constant water-to-cement ratio, concrete incorporating Ha Nam limestone exhibited comparable compressive strength to that of Hoa Binh basalt, yet demonstrated a higher elastic modulus. Based on the experimental data, this study proposes using preliminary coefficients of 21.000 and 19.700 - corresponding to Ha Nam limestone and Hoa Binh basalt aggregates, respectively - for estimating the modulus of elasticity in accordance with EN 1992-1-1:2004. For estimations in accordance with ACI CODE-318-25, the recommended preliminary factors are 5.250 and 4.810, respectively.

Vehicle and time specific crash modelling on selected rural highway curves using geometric and speed parameters: A transformed linear regression approach

2025-10-28T09:13:52+00:00

This study develops vehicle and time-specific crash rate prediction models for rural highway curves using high-resolution geometric and speed data. A 30km segment of State Highway-1 in Karnataka, India, encompassing 32 horizontal curves, served as the study site. Detailed data collection included 10 years of crash records, traffic volume count, LiDAR-based geometric features, and spot speeds recorded from laser speed cameras. Distinct models were built for motorized two-wheelers (MTW), passenger cars (CAR), heavy commercial vehicles (HCV), and for both daytime and nighttime conditions. The study offers a novel contribution by incorporating nighttime crash rate modelling rarely addressed due to challenges in data availability, and by developing disaggregated models for multiple vehicle classes. A backward stepwise regression (BSR) approach with square root transformation was employed, ensuring model transparency and interpretability. Sight-distance deficiency consistently emerged as the most influential predictor of crash rate, highlighting the critical role of visibility on curved segments. Validation through Leave One Out Cross Validation (LOOCV) confirmed acceptable predictive performance (R² = 0.43-0.80), with residuals exhibiting normal distribution. The findings underscore the importance of curve geometry and visibility in crash risk and provide actionable insights for design audits and safety interventions on rural highways.

Landslide detection and susceptibility analysis: A case study in Pieng stream catchment, Son La province

2025-12-18T02:36:04+00:00

Landslides are a major natural hazard causing significant property damage and loss of life worldwide. In this study, an enhanced landslide inventory was developed for the Pieng Stream catchment (Son La Province) using Object-Based Image Analysis (OBIA) combined with field surveys. Twelve conditioning factors were used to model landslide susceptibility through four machine learning algorithms: extreme gradient boosting (XGB), random forest (RF), multi-layer perceptron (MLP) and logistic regression (LR). Additionally, model interpretation was supported by SHAP, MDA, and PDP analyses. The results demonstrated a high level of reliability for the OBIA method (TPR = 0.886, TS = 0.602). Among the tested models, the XGB model showed the best performance, achieving an AUC of 0.961, an F1 score of 0.915, and an accuracy of 0.915 on the testing dataset. The two most influential predictors identified were lineament density and aspect. An increase in landslide probability was observed with increasing slope, relative relief, lineament density, river density and aspect (0-150°). A total of 88% of testing landslide points were correctly classified within high to very high susceptibility areas, while areas outside the AOA covered merely 0.79% of the study region, indicating a high level of model applicability.

Sliding mode control of a rotary double parallel inverted pendulum

2025-10-26T14:14:50+00:00

This study explores the stabilization of a rotating parallel inverted pendulum, a nonlinear, underactuated system challenging conventional control methods. Unlike prior work focusing on linear techniques like Linear Quadratic Regulator, which lack robustness to uncertainties, this research introduces a Sliding Mode Control (SMC) strategy. Utilizing a linear sliding surface, the proposed SMC ensures finite-time convergence and robust performance despite model uncertainties and disturbances. Numerical simulations and experimental results validate the controller’s ability to stabilize the pendulum and rotating base, demonstrating its superiority over traditional methods and potential applicability to similar complex mechanical systems.

Dynamic Reliability Thresholds and Adaptive Load Management for Safety-Optimized Ageing Concrete Bridges

2025-11-20T08:44:07+00:00

This study presents a reliability-based framework for enhancing the durability design of ageing concrete bridges by integrating time-dependent reliability (TDR) and performance-based durability design (PBDD) principles. The proposed methodology addresses these gaps through: (1) probabilistic modeling of resistance degradation and load variations, (2) dynamic adjustment of reliability thresholds to optimize maintenance strategies, and (3) a systematic linkage between target reliability indices and operational constraints. Case studies demonstrate how real-time load management can extend service life while meeting international safety standards. The framework bridges theoretical advancements with practical applications, offering actionable insights for engineers to mitigate lifecycle costs and structural risks.

High-performance geopolymer concrete: a review of recent developments

2025-11-21T07:44:02+00:00

High-performance geopolymer concrete (HPGC) is an advanced construction material designed for enhanced strength, sustainability, and cost-effectiveness. Based on aluminosilicate precursors activated by alkaline solutions, HPGC offers notable advantages over conventional Portland cement concrete. This review synthesizes recent global and Vietnamese studies on HPGC, addressing its constituent materials, mix design, fresh and mechanical properties, durability, environmental impact assessment, microstructure, and potential applications. Industrial by-products such as fly ash, slag, rice husk ash, silica fume, and metakaolin are commonly used in various proportions with alkali activators. HPGC typically shows greater workability, viscosity, and cohesion than traditional concrete. Its compressive strength ranges from 50 to 91 MPa and can exceed 130 MPa under thermal curing, with rapid strength gain supporting faster construction. In terms of durability, HPGC exhibits low permeability and strong resistance to acid, sulfate, and steel corrosion, especially in marine environments. It also offers a significant reduction in carbon emissions compared to conventional concrete. In Vietnam, research on HPGC is still limited but shows promising potential for marine infrastructure applications.

Durability of reinforced geopolymer concrete using coral aggregates and seawater under accelerated corrosion

2025-09-25T08:26:50+00:00

This study evaluates the corrosion resistance of geopolymer concrete made with coral sand, coral rock, and seawater, materials abundant on offshore islands. Three mixtures were designed using varying proportions of fly ash and blast furnace slag as binders. Compressive strengths after 28 days were 20.11 MPa, 25.32 MPa, and 30.15 MPa for GPS-1, GPS-2, and GPS-3, respectively. Accelerated corrosion tests (NT Build 356) revealed that higher strength samples exhibited lower average current intensity before cracking (5.49 mA for GPS-1 vs. 4.68 mA for GPS-3) and longer times to cracking (204 h vs. 324 h). Chloride ion concentrations near the steel at the onset of cracking were similar across mixtures (0.53 % – 0.58 % by dry concrete weight), but initial chloride contents were lower in high-strength samples (0.08 % in GPS-3 vs. 0.22 % in GPS-1), indicating better resistance to ion penetration. Microstructural analyses (SEM, FTIR, XRD) confirmed that GPS-3 had a denser matrix, more developed geopolymer gel, and reduced presence of unreacted crystalline phases. These results demonstrate that optimizing binder content and microstructure enhances both mechanical performance and corrosion resistance, making geopolymer concrete with marine-derived materials a viable solution for durable coastal infrastructure.

Vehicle Classification Using Combined Laser Rangefinder and Pyroelectric Infrared Sensors in a Real Dynamic Environment

2025-12-11T08:34:22+00:00

This paper presents a vehicle classification approach for real-world dynamic environments based on sensor fusion between a laser rangefinder (LRF) and a pyroelectric infrared (PIR) sensor. By integrating geometric shape information from the LRF with thermal distribution patterns captured by the PIR sensor, the system extracts distinctive features that effectively suppress noise introduced by external environmental variations. A lightweight neural network is developed for classification, achieving a minimum accuracy of 91% for specific vehicle types and an average accuracy of 94% across all categories. Owing to its high accuracy and low computational cost, the proposed model is well-suited for implementation in portable embedded platforms, functioning as intelligent measurement nodes within Intelligent Transportation Systems (ITS).

Design and Inverse Kinematics of Continuum Robots

2025-12-01T14:04:05+00:00

Continuum robots, characterized by their hyper-redundant and flexible structures, have gained significant attention in fields such as minimally invasive surgery, remote inspection, and soft manipulation. Their complex design and highly nonlinear kinematic behavior present substantial modeling and control challenges. This paper presents a complete framework encompassing the design, modeling, and inverse kinematics solution for a novel two-segment, tendon-driven continuum robot utilizing an elastic spring backbone for enhanced compliance and structural simplicity. A constant curvature forward kinematic model is presented. Subsequently, an efficient numerical approach for solving the challenging inverse kinematics problem is introduced by adapting the Jacobian-based Newton-Raphson method and incorporating the Moore-Penrose Pseudoinverse. This strategy effectively manages the robot's redundancy, ensuring smooth and reliable trajectory generation. Experimental verification confirms the robot's feasibility, demonstrating that the system successfully navigates along the trajectories computed by the inverse kinematics, thus validating the reasonableness of the kinematic model and ensuring seamless and smooth operation. These findings provide a robust foundation for improving motion planning of simple continuum robot platforms in practical applications.