The passage of trains by railway noise barriers induces vibrations that may affect their fatigue performance and reduce their service life. However, long-term field monitoring of noise barriers under complex environmental and operation conditions remains rare. This study develops an interpretable machine learning (ML) framework to investigate the aerodynamic pressure and dynamic behaviors of noise barriers based on a nine-month long-term field monitoring campaign, yielding 12810 train runs over 105 valid days. Input variables include train type, speed, temperature, wind speed and direction, relative humidity, and air pressure, while the target responses cover train-induced aerodynamic pressure, stress near the base of the steel post, and displacement at the post top. Eight ML models, including four traditional and four ensemble algorithms, were used and systematically compared to evaluate their predictive capabilities and robustness. Ensemble models, particularly Gradient Boosting Decision Tree (GBDT), Light Gradient Boosting Machine (LGBM), and Extreme Gradient Boosting (XGBoost), achieved the best predictive performance, with R² values exceeding 0.935 for stress and displacement, and 0.895 for pressure. XGBoost, offering a strong balance of predictive accuracy and computational efficiency, was selected for SHapley Additive exPlanations (SHAP)-based interpretability analysis to uncover the physical relationships behind the data-driven predictions. Results reveal that aerodynamic pressure was the most challenging response to predict, given its higher sensitivity to turbulent airflow and environmental fluctuations, whereas stress and displacement exhibited more stable and predictable patterns. SHAP analysis identified train speed and type as the most influential factors across all responses. While environmental factors had comparatively lower influence, temperature and instantaneous wind direction consistently showed higher importance among them. Relative humidity has a moderate effect on aerodynamic pressure but a minor impact on dynamic behavior. Air pressure and wind speed exhibit limited influence on all outputs. These findings highlight the novelty and effectiveness of integrating long-term monitoring data, ML methods, and SHAP-based interpretability, offering new insights into the dynamic behavior of railway noise barriers.

Reliability analysis is crucial for evaluating structural performance, and advancements in computational technologies have enhanced the application of Finite Element Modeling (FEM) in this field. However, existing reliability methods face challenges, particularly with approximation methods that struggle with implicit limit state functions and simulation methods with cost. Metamodel-based approaches have gained popularity for their balance of efficiency and accuracy, yet most algorithms focus on independent variables. This paper presents an improved metamodel-based algorithm designed to assess structural reliability while considering variable dependency and FEM. The algorithm provides a methodology that ensures a rapid convergence of the iterative process when building the metamodel in the region of interest, i.e. where the Most Probable Point (MPP) is located. It uses a modified Design of Experiments (DoE) and integrates copula theory to model the joint probability distribution function (PDF). The algorithm enriches the experimental points matrix by finding points close to the Limit State Function (LSF) at 0 with high prediction variance using learning functions and with high joint PDF values using copula theory. Validation through various examples and an application example of a reinforced concrete bridge indicates that the proposed methodology is more efficient than most recent algorithms without reducing accuracy.

Load capacity assessment of concrete dams often includes verification of the stability for multiple separate failure modes, such as sliding and overturning. However, in the case of dams, the underlying failure mechanism for these failure modes may be too idealized, and the analysis could yield inaccurate results. Previous research has, for example, shown that regular rigid-body sliding failure analysis provides inaccurate load capacity estimates for dams with uneven interface geometries. This article discusses the behavior of such dams and presents a failure mode that combines the traditional sliding and overturning failures. The failure mode is termed combined sliding and overturning and serves as an intermediate to the traditional failure modes. It allows for the assessment of concrete dams with uneven interface geometries, whose behavior is not expected to be fully represented by only sliding or overturning. To estimate the load capacity for the presented failure mode, an analytical formulation based on simple force and moment equilibrium is provided. The formulation is compared with finite element simulations and previously reported results from experimental scale model tests and is shown to accurately predict the load capacity.

Noise barriers play a crucial role in mitigating railway noise, with the aerodynamic pressure exerted by passing trains being a key factor in their structural design, particularly for those installed along high-speed railways. While previous studies have focused on the effects of train speed, geometry, and distance from the track centre, and have developed models incorporating these factors, limited attention has been given to the impact of bilateral layouts and barrier height on this pressure. Quantitative assessments of these two factors remain scarce, and existing pressure calculation models inadequately address their influence. This study addressed these gaps by employing computational fluid dynamics (CFD) simulations, validated by field test data, to qualitatively and quantitatively analyze the effects of barrier layout and height on the aerodynamic pressure acting on vertical noise barriers. The results demonstrate that two distinct transient pressure fluctuations over time are generated by the train’s nose and tail, in agreement with the findings of the field tests. A bilateral layout increases peak pressure by up to 8.5%, particularly as the distance to the train centreline decreases. Moreover, increasing barrier height from 2 to 4 m resulted in a maximum pressure amplification of 13.23%, though the amplification rate diminished with further height increases. To address the limitations of existing pressure calculation models, an exponential model was developed to account for the amplification effect of bilateral layouts, while a logarithmic correction factor was introduced to account for barrier height. These models were integrated into a comprehensive aerodynamic pressure calculation framework, effectively capturing the combined impacts of barrier layout and height. Validated through simulations, the proposed model offers a more accurate and practical approach for predicting train-induced aerodynamic pressure on noise barriers, providing valuable insights to inform their structural design.

The use of digital twin (DT) technology within the engineering and construction (E&C) industry is valuable for practical applications in asset management of structures. Functional DT in E&C, however, are still in initial stages of development. Efforts towards standardization of concepts and procedures are necessary to build on existing knowledge and drive progress further on functional DT. This paper proposes a DT of a snow gallery, part of the Iron Ore railway in northern Sweden. The gallery was instrumented with a structural health monitoring (SHM) system that feeds data in real time to the DT, which also includes a 3D model of the gallery. The proposed methodology can be replicated to different structures and scaled for larger amounts of data. The SHM data and the 3D digital model of the snow gallery are connected in a single, integrated platform that enables improved decision-making for maintenance of the gallery. To promote clarity and progress within the field, the proposed DT’s maturity level is classified in terms of autonomy, intelligence, learning, and fidelity. The snow galleries, the SHM system, and the proposed DT are all presented and discussed, following a brief review on DT, the importance of level classification, and predictive maintenance.

Railway noise barriers are an essential piece of infrastructure for reducing noise propagation. However, these barriers experience aerodynamic loads generated by high-speed trains, leading to dynamic effects that may compromise their fatigue capacity. The most common structural design for railway noise barriers consists of vertical configurations of posts and panels. However, there have been few dynamic analyses of steel post/wood panel noise barriers under train-induced aerodynamic loads. This study used dynamic finite element analysis to assess the dynamic behavior of such noise barriers. Analysis of a 40-m-long noise barrier model and a triangular simplified load model, the latter of which effectively represented the detailed aerodynamic load, were first used to establish the model and input of the moving load during dynamic simulation. Then, the effects of different parameters on the dynamic response of the noise barrier were evaluated, including the damping ratio, the profile of the steel post, the span length of the panel, the barrier height, and the train speed. Gray relational analysis indicated that barrier height exhibited the highest correlations with the dynamic responses, followed by train speed, post profile, span length, and damping ratio. A reduction in the natural frequency and an increase in the train speed result in a higher peak response and more pronounced fluctuations between the nose and tail waves. The dynamic amplification factor (DAF) was found to be related to both the natural frequency and train speed. A model was proposed showing that the DAF significantly increases as the square of the natural frequency decreases and the cube of the train speed rises.

Strengthening existing masonry arch structures with externally bonded composite systems has become a common technique to enhance structural performance while allowing continued use of the infrastructure during retrofitting. Among these systems, textile-reinforced mortar (TRM) solutions applied to the intrados of arches have shown promising results. The effectiveness of mortar-based materials, such as fabric reinforced cementitious matrix (FRCM), TRM, and steel reinforced grout (SRG), relies heavily on the interaction between the matrix, the masonry substrate, and the embedded fibres. However, when applied to concave masonry surfaces, the curvature introduces additional normal stresses that may hinder stress transmission and bond performance. This study investigates how substrate curvature and the use of spike anchors influence the failure mechanisms in FRCM strengthened masonry. Single shear-lap tests were conducted on curved masonry specimens to analyse load transfer behavior and identify failure modes. The results reveal that curvature significantly alters typical failure mechanisms observed on flat substrates, even leading to the emergence of subcategories within standard failure modes. Moreover, the incorporation of spike anchors was found to locally modify failure progression, improving overall bond performance. The paper presents a detailed classification of observed failure modes, supported by histograms that quantify the effect of curvature and the spike anchors. These findings provide valuable insight into the design of strengthening systems for curved masonry elements and highlight the importance of tailored anchorage strategies.

Structural Health Monitoring (SHM) is crucial for ensuring bridge safety, yet many methods rely on baseline data or known damage states—often unavailable for aging structures. To address this, we propose a new approach that combines Artificial Neural Networks (ANNs) with matrix profiling (MP) to create a “pseudo-baseline” for predicting bridge behavior. Physics-Informed Neural Networks (PINNs) incorporate physical laws into the model, while MP detects patterns and subtle anomalies in structural data. This method links structural responses, like strain and displacement, to environmental factors such as temperature and humidity. By analyzing these relationships, we can model normal bridge behavior without needing complete historical data. The approach is validated using performance metrics such as R2, Root Mean Square Error (RMSE), and residual analysis. Our combined method offers an innovative solution for real-time anomaly detection, providing a more accurate and proactive tool for long-term bridge monitoring.

Reinforced concrete (RC) trough bridges form a crucial part of Europe's railway infrastructure. These structures consist of a U-shaped cross-section (two longitudinal beams and a slab) designed to accommodate the ballast. For the case of the Iron Ore Line, a critical corridor located in the north of the country, RC trough bridges represent about 40 % of the railway bridge population. Many of these bridges have surpassed 50 years of service, enduring over 10 million cycles of fatigue loading, with increases of axle loads since their construction due to demands associated with the iron ore extraction and transportation. As these structures approach critical maintenance or replacement decisions, understanding their long-term performance and remaining capacity is essential. This study experimentally investigates the degradation behavior of a full-scale RC trough bridge subjected to progressive cyclic loading, simulating fatigue effects over time. During the controlled laboratory tests, the overall performance of the bridge is assessed, focusing on the stiffness loss of slab and beams, cracking, and force redistribution. Fatigue verifications based on Eurocode EN1992–1–1 and fib Model Code 2020 are performed alongside the reliability analysis using the First-Order Reliability Method (FORM) to evaluate structural safety levels. While fatigue damage is evident in the slab's tensile zone, the overall structural response indicates that the bridge maintains its functional capacity after simulating 48 years of service. However, the code-based reliability analysis indicates lower-than-target reliability levels for reinforcement, suggesting a conservative estimation of reinforcement capacity to withstand cyclic loading.

This paper investigates the current landscape of multiscale studies in concrete composites incorporating molecular dynamics (MD) methods. Through a thorough literature analysis, it was determined that finite element, discrete element, homogenization, microphysical characterization, and machine learning methods are better suited for integration with MD in multiscale studies of concrete composites. The paper delves into MD's application characteristics and the selection of force fields in multiscale studies and provides a summary of the combined applications between MD and various methods. Challenges identified include the optimization of MD simulations and the appropriate selection of combined methods. The conclusions underscore the growing recognition of MD's significance, advocating for rational multi-method integration in multiscale approaches to effectively advance research on concrete composites.