Conventional fatigue assessment of steel railway bridges relies on strain monitoring and cycle counting (e.g., rainflow) to obtain stress spectra. Motivated by accelerometers’ long-term stability and ease of installation, recent studies seek to infer local stress histories from bridge vibrations; however, this remains challenging because acceleration signals are noisy and confounded by structural dynamics, variable loading, resonance, and train–track interactions. To address these challenges, this study introduces Vib2StressNet, a deep-learning architecture that maps multi-channel vertical vibration signals directly to fatigue stress spectra. Unlike sequence-to-sequence models that reconstruct full time histories, this approach bypasses intermediate steps to focus on damage accumulation. The architecture integrates convolutional layers with a multi-head self-attention mechanism that captures information across multiple frequency bands. A critical design feature is the inclusion of train speed as an auxiliary scalar embedding, allowing the network to dynamically adapt to speed-dependent resonance. Vib2StressNet demonstrated strong generalizability across three Swedish railway bridges with diverse structural configurations, train types, and dynamic loading conditions. Under resonance and variable-speed conditions, adding train speed significantly improved prediction accuracy, reducing the mean squared error (MSE) by more than 20%. Relative to a previous sequence-to-sequence baseline model that reconstructs stresses before rainflow counting, Vib2StressNet simplifies the workflow and reduces MSE from 59.8 to 0.34 with more test samples. Interpretability analyses further confirm that preserving high-frequency components associated with axle impacts improves accuracy. These findings establish Vib2StressNet as a cost-effective tool for long-term monitoring that supports predictive maintenance without permanent strain instrumentation.

Modal properties are widely used for structural model updating and damage detection, yet their sensitivity to environmental conditions, particularly temperature can complicate interpretation. This paper investigates how seasonal temperature variations affect structural identification based on modal properties, using a falsification-based approach. A railway bridge in service is used as a case study, with data from both forced vibration testing and long-term monitoring. Freezing temperatures are found to increase overall stiffness by up to 90%, including a significant rise in boundary stiffness, an increase in ballast modulus of up to 10 GPa, and the activation of rail continuity stiffness. The set of plausible physical models shifts between winter and summer, demonstrate that structural behaviour varies seasonally and that identification results are not directly transferable across temperature ranges. Compared to traditional residual minimisation, model falsification provides a more comprehensive and robust set of plausible models and avoids unreliable parameter estimates. These findings underscore the need for temperature-sensitive identification strategies and adaptive models to support accurate interpretation, damage detection, and seasonally informed maintenance decisions in structural health monitoring.

In all civil engineering practices, the goal is to produce effective and safe structures, minimising material usage while conforming to all relevant safety and comfort criteria. To achieve materially efficient bridges, it is of the utmost importance to investigate and optimise various design alternatives and compare them. In this paper, a vast array of simply supported steel-concrete composite railway bridges for high speed railway traffic are optimised using a Genetic Algorithm. Different cross-sections, steel grades, and span lengths are analysed to achieve the lowest possible steel mass. Four different cross-sections are examined: double I-girders, double box girders, a semi box girder with a joint lower flange, and a box girder with inclined webs. Steel grades from S235 to S460 and span lengths from 20-70 m are used in the optimisation. As constraints, Eurocode criteria for accelerations, the ultimate, and serviceability limit states, fatigue, beam end rotations, and beam deflections are used. Web slenderness is limited to avoid web breathing. The results indicate that I-girders produce the cross-section with the lowest mass. The most decisive criteria are accelerations for span lengths below about 50 m and the serviceability limit state stress check above about 50 m. In all cases, cross-sections with very slender webs render the lowest steel mass.

Effective anomaly detection and classification in bridge monitoring data critically depend on robust feature extraction to ensure reliable performance across diverse sensors, loading scenarios, and anomaly types. This study presents a novel deep metric learning-based framework that automates feature extraction, addressing the limitations of traditional manual methods that often lack generalization. The proposed approach employs continuous wavelet transforms to convert time-series signals into time-frequency representations, followed by a convolutional neural network trained with metric learning to produce low-dimensional feature optimized for anomaly classification. These embeddings capture similarities between anomalies, enabling the model to distinguish normal loading events from anomalous ones while classifying specific anomaly types. The method was validated on bridge monitoring data collected from strain gauges and accelerometers under varying loading scenarios, achieving high detection and classification accuracy. This robust and scalable framework highlights the potential of deep metric learning to advance automated bridge monitoring.

To reduce investment cost and environmental impact of new civil engineering structures, an obvious approach is to reduce material usage. For bridge structures this poses a challenge, as demands put on the structure differ when analysing static and dynamic loading, with the end result being that bridge design is an iterative task. To address the problem, an optimisation procedure using a Genetic Algorithm (GA) is employed on a single span steel-concrete composite railway bridge, the Bryngeå Bridge on Botniabanan in northern Sweden. The optimisation is performed to obtain a lighter structure while still conforming to demands put on bridge structures presented in the Eurocodes. The criteria used in the work are quasistatic design calculations for the ultimate limit state, serviceability limit state and fatigue limit state, as well as a dynamic evaluation performed for train speeds up to 300 km/h, using the High Speed Load Models in the Eurocodes. In total, two optimisations are performed, one including retrofitted dampers on the bridge, and one exluding the dampers. The work is limited to minimising the mass of the steel girders in the bridge. Using accurate material parameters obtained from previous system identification of the bridge, results show that accelerations in conjunction with ultimate limit state loading limit the design of the bridge, and that I-girders with large tension flanges and small compression flanges produce the lightest cross-sections. Furthermore, results indicate that GA is an efficient tool for obtaining lighter structures.

Säkerhetskontroller av konstruktioner i den byggda miljön bygger på det semiprobabilistiska säkerhetsformatet som kallas partialkoefficientmetoden. Med specifika koefficienter för lasteffekter och bärförmågor bedöms säkerheten mot ett givet gränsvärde. Principen bakom metoden är att partialkoefficienterna ska beakta respektive delvariabels osäkerhet. Formatet är relevant om alla koefficienterna bestäms tillsammans, endast då uppnås avsedd säkerhet.

Bärighetsberäkning av befintliga konstruktioner bygger på samma säkerhetsformat som dimensioneringen av nya konstruktioner. Det finns dock anledning att differentiera kontrollerna för att beakta befintliga konstruktioners särskilda egenskaper, som till exempel att konstruktionerna redan är byggda, att lasterna som de utsätts för är kända, att objektspecifik data kan samlas in, och dessutom, att åtgärder på befintliga konstruktioner typiskt är kostsamma och resurskrävande.

Denna rapport redovisar en historisk genomgång av föreskrifterna för bärighetsberäkningar och några av de bakomliggande arbetena. Jämförande beräkningar görs för en av de första föreskrifterna från 1996 och dagens gällande TRVINFRA-00331. Dessutom sammanfattas några internationella föreskrifter för bärighetsberäkningar.

Den sannolikhetsteoretiska bakgrunden till partialkoefficientmetoden gås igenom och kompletta probabilistiska känslighetsanalyser har genomförts för att visa vilka förutsättningar som är avgörande för partialkoefficienterna. Stora mängder data för tågtrafiklaster från detektormätningar och vägtrafiklaster från BWIM-mätningar har beaktats i analyserna. Statistiska data för andra variabler har hämtats från litteraturen.

Resultaten visar på signifikanta skillnader mellan modellerna för bärförmåga, med betydande osäkerheter i t.ex. betongs materialegenskaper vilket blir dominerande vid bestämningen av partialkoefficienter. För armering och konstruktionsstål blir istället trafiklastens osäkerhet dominerande. Denna skiftning av dominerande variabler försvårar enhetliga val av koefficienter och motiverar differentiering mellan såväl brottmoder som typer av trafiklast.

Resultaten visar på stor potential med en kalibrering av säkerhetsformatet med syftet att bedöma befintliga broar mot en relevant säkerhetsnivå, vilket i förlängningen bör leda till bättre beslut om att behålla, reparera eller byta ut broar. Denna rapport redogör för hur en sådan kalibrering kan genomföras och redovisar underlag för de ingående stokastiska variablerna.

Accurate differentiation between damage-related anomalies and data errors is a critical challenge in bridge monitoring. This paper presents a data-driven framework for anomaly detection and classification, addressing the question: How can anomalies be classified in multi-sensor bridge monitoring to distinguish structural changes from noise? The framework combines an adapted anomaly taxonomy with a deep neural network trained on synthetic data. It is validated using long-term monitoring data from a railway bridge, incorporating strain gauges, accelerometers, and an inclinometer. In offline training, the model achieves high precision, recall, and F1-scores, effectively detecting anomaly classes across sensor types. For online prediction, it provides anomaly type percentages and visualizations over daily, weekly, and annual timeframes, distinguishing frequent noise-related anomalies from rare anomalies signaling structural changes. Requiring one month of training data, the framework delivers a scalable solution for bridge monitoring and lays the groundwork for future self-learning anomaly detection in infrastructure management.

This study introduces a methodology that integrates two complementary approaches to estimate stress at various locations on a bridge using sparse vibration data. The approach is validated with in-service monitoring data collected from a composite railway bridge in Sweden. Specifically, a deep learning sequence model is used to predict stress histories based on acceleration measurements from multiple locations. Following this, the local response function method is applied, using localized models of the bridge’s fatigue critical details to enable precise stress predictions in critical areas. By combining these techniques, the methodology achieves conservative predictions of stress ranges and cycles at instrumented locations, reducing the reliance on extensive instrumentation and offering a more efficient solution for structural health monitoring.

The bridge network is progressively aging, with an alarming proportion of bridges over 100 years. This situation engenders substantial risks to the overall reliability of transportation networks, requiring innovative methods for efficient management. Monitoring can provide a direct source of information about structural behavior generating alerts when changes occur. Real-time alerts enable effective infrastructure management and decision-making during damage or anomalous situations. However, monitoring can result in a large amount of data that is often difficult to convert into valuable information in real time. This paper presents an approach for real-time detection of abrupt damage occurrence in bridges using unsupervised anomaly detection algorithms and strain/acceleration measurements. The approach incorporates the separation of measurements into events having the same loading nature and the construction of three feature matrices based on statistical features, time-frequency features, and wavelet spectrum features. It includes the evaluation of five anomaly detection algorithms including Isolation Forest, One-Class Support Vector Machine, Robust Random Cut Forest, Local Outlier Factor, and Mahalanobis Distance. The approach is illustrated with a case study of a steel-bascule-railway bridge, that has experienced a brittle cracking event during monitoring. Results highlight the robustness of One-Class Support Vector Machine, Isolation Forest, and Local Outlier Factor algorithms in promptly detecting abrupt changes across different features. The separation of strain and acceleration data into loading-based events, coupled with the comparison of previous and new event features, provides robust feature matrices for effective damage detection. Enhanced detection and higher scores are particularly attributed to time-frequency domain features during damage occurrence. The presented approach can be used as a base on how to perform real-time anomaly detection within the context of bridge monitoring.

The trend in recent years to build longer and slender bridges has created new challenges in terms of safety and structural integrity. This study focuses on the effects of windinduced vibrations on the hangers of the Bomarsund Bridge in Åland, Finland. Those close to the mid-span have exhibited significant vibrations, requiring a thorough investigation to understand the response of the bridge to wind action. Computational Fluid Dynamics (CFD) simulations were performed using ANSYS Fluent to calculate the aerodynamic coefficients characterizing the given section (i.e. Strouhal number, drag, shedding frequencies, etc.) and to investigate the vortex-induced vibration (VIV) phenomena at different wind speeds. The results showed inconsistent drag coefficients at varying wind speeds and lower drag for geometries with rounded edges. The study highlighted the significant dependence of the Strouhal number on wind speed, challenging traditional geometry-based estimations. These findings can be used to implement effective measures to mitigate wind-induced vibrations.