A realistic and economical dynamic assessment of railway bridges requires input parameters that correspond to reality. In this context, the applied damping properties of the structure have a decisive influence on the results in the prediction of resonance effects and further in the assessment of the compatibility between rolling stock and railway bridges. The standard prescribes damping factors depending on the type of structure and the span to be used in dynamic calculations. However, these factors can be regarded as very conservative values which do not represent reality. Thus, in situ measurements on the structure are often necessary to classify a bridge categorised as critical in prior dynamic calculations as non-critical. Regarding in situ tests, a measurement-based determination of the damping factor is inevitably accompanied by a scattering of the generated results due to the measurement method used and as a result of the individual scope of action of the test-evaluating person and this person’s interpretation of the measurement data. This paper presents novel evaluation methods and analysis tools for determining the damping factor based on measurements in the frequency and time domains, intending to reduce the scatter of the results and limit the scope of action of the person evaluating the test. The main aim is to provide simple and easy-to-use evaluation algorithms for practical applications without additional data transformations and to define clear principles of action for the data-based evaluation of realistic and high damping factors. Based on in situ tests on 15 existing railway bridges, the data-based procedure for determining the damping factor is explained, and the methods are compared in the time and frequency domains. It is shown that a clearly defined evaluation algorithm can significantly reduce the scattering of results. Furthermore, it is demonstrated that forced vibration excitation and evaluation in the frequency domain provide the best results in reliable, reproducible, and high damping factors.

Railway systems are critical components of transportation networks requiring consistent maintenance. This paper proposes a novel data-driven approach to detect various maintenance needs of railway track systems using acceleration data obtained from a passenger train in operation. The framework contains four modules. Firstly, data pre-processing and cleansing are performed to extract useful data from the whole dataset. Then, condition-sensitive features are extracted from the raw data in three different domains of time, frequency, and time–frequency. In the third module, the best subset of measurement features that characterize the state of the tracks are selected using the analysis of variance (ANOVA) algorithm which eliminates irrelevant characteristics from the feature set of responses. Finally, a multilabel classification algorithm based on the cascade feed-forward neural network (CFNN) is used to classify the type of maintenance needs of the track. An open-access dataset from a field study in Pennsylvania, USA, is used in this study for validation of the proposed method. The results indicate that employing a CFNN can achieve 95% accuracy in identifying two maintenance activities, tamping and surfacing, using time-domain features. Moreover, an extensive analysis has been conducted to evaluate the influence of various feature extraction and selection methods, diverse classification algorithms, and different types of accelerometers (uni-axial and tri-axial) on the accuracy of the proposed method.

Imbalanced loads in freight railway vehicles pose significant risks to vehicle running safety as well as track integrity, increasing the likelihood of derailments and increasing track wear rate. This study presents a robust machine learning-based methodology designed to detect and classify transverse imbalances in freight vehicles using dynamic rail responses. The proposed approach employs wayside monitoring systems with accelerometers and strain gauges, integrating advanced feature extraction methods, including principal component analysis, log-mel spectrograms, and multi-feature-based techniques. The methodology enhances detection accuracy by normalizing features to eliminate environmental and operational variations and employing data fusion for sensitive index creation. It is capable of distinguishing between different severity levels of imbalanced loads across various wagon types. By simulating scenarios with typical European freight wagons, the study demonstrates the effectiveness of the approach, offering a valuable tool for railway infrastructure managers to mitigate risks associated with imbalanced loads. This research contributes to the field by providing a scalable, non-invasive solution for real-time monitoring and safety enhancement in freight rail operations.

High-speed railway bridges are essential components of any railway transportation system that should keep adequate levels of serviceability and safety. In this context, drive-by methodologies have emerged as a feasible and cost-effective monitoring solution for detecting damage on railway bridges while minimizing train operation interruptions. Moreover, integrating advanced sensor technologies and machine learning algorithms has significantly enhanced structural health monitoring (SHM) for bridges. Despite being increasingly used in traditional SHM applications, studies using autoencoders within drive-by methodologies are rare, especially in the railway field. This study presents a novel approach for drive-by damage detection in HSR bridges. The methodology relies on acceleration records collected from multiple bridge crossings by an operational train equipped with onboard sensors. Log-Mel spectrogram features derived from the acceleration records are used together with sparse autoencoders for computing statistical distribution-based damage indexes. Numerical simulations were performed on a 3D vehicle–track–bridge interaction system model implemented in Matlab to evaluate the robustness and effectiveness of the proposed approach, considering several damage scenarios, vehicle speeds, and environmental and operational variations, such as multiple track irregularities and varying measurement noise. The results show that the proposed approach can successfully detect damages, as well as characterize their severity, especially for very early-stage damages. This demonstrates the high potential of applying Mel-frequency damage-sensitive features associated with machine learning algorithms in the drive-by condition assessment of high-speed railway bridges.

Supervised learning classification has arisen as a powerful tool to perform data-driven fault diagnosis in dynamical systems, achieving astonishing results. This approach assumes the availability of extensive, diverse and labeled data corpora for training. However, in some applications it may be difficult or not feasible to obtain a large and balanced dataset including enough representative instances of the fault behaviors of interest. This fact leads to the issues of data scarcity and class imbalance, greatly affecting the performance of supervised learning classifiers. Datasets from railway systems are usually both, scarce and imbalanced, turning supervised learning-based fault diagnosis into a highly challenging task. This article addresses time-series data augmentation for fault diagnosis purposes and presents two application cases in the context of railway track. The case studies employ generative adversarial networks (GAN) schemes to produce realistic synthetic samples of geometrical and structural track defects. The goal is to generate samples that enhance fault diagnosis performance; therefore, major attention was paid not only in the generation process, but also in the synthesis quality assessment, to guarantee the suitability of the samples for training of supervised learning classification models. In the first application, a convolutional classifier achieved a test accuracy of 87.5% for the train on synthetic, test on real (TSTR) scenario, while, in the second application, a fully-connected classifier achieved 96.18% in test accuracy for TSTR. The results indicate that the proposed augmentation approach produces samples having equivalent statistical characteristics and leading to a similar classification behavior as real data.

This study explores the feasibility of electromagnetic acoustic transducers (EMATs) for ultrasonic rail inspection, focusing on bulk wave generation from the rail head and on defect detection at the central part of the rail foot. As a contactless method, EMATs could overcome some known limitations of conventional ultrasonic techniques, but require further validation. Different campaigns of experimental tests were performed, evaluating, by means of a probability of detection approach, the response of the technique to several artificial semi-elliptical flaws of increasing size and by considering two sensors characterized by different working frequencies. In contact, static tests allowed to assess the basic feasibility of the inspection technique and showed a linear response to defect size, saturating when defect width exceeded the rail web thickness. Dynamic tests allowed to introduce the effects of lift-off on signal responses. During all tests, the higher-frequency sensor outperformed the lower-frequency one. Finally, full-scale bogie tests on an indoor permanent track installation, comprehensive of defective rails, confirmed the higher flaw detection rates of the higher-frequency sensor, with minimal detection failures despite occasional false alarms. EMATs showed encouraging results for in-motion rail inspection: with further technical development and optimization, this technique could enhance ultrasonic rail inspection by diagnostic trains.

Drive-by techniques for bridge health monitoring have drawn increasing attention from researchers and practitioners, in the attempt to make bridge condition-based monitoring more cost-efficient. In this work, the authors propose a drive-by approach that takes advantage from bogie vertical accelerations to assess bridge health status. To do so, continuous wavelet transform is combined with multiple sparse autoencoders that allow for damage detection and localization across bridge span. According to authors’ best knowledge, this is the first case in which an unsupervised technique, which relies on the use of sparse autoencoders, is used to localize damages. The bridge considered in this work is a Warren steel truss bridge, whose finite element model is referred to an actual structure, belonging to the Italian railway line. To investigate damage detection and localization performances, different operational variables are accounted for: train weight, forward speed and track irregularity evolution in time. Two configurations for the virtual measuring channels were investigated: as a result, better performances were obtained by exploiting the vertical accelerations of both the bogies of the leading coach instead of using only one single acceleration signal.

Accurate monitoring of track irregularities is very helpful to improving the vehicle operation quality and to formulating appropriate track maintenance strategies. Existing methods have the problem that they rely on complex signal processing algorithms and lack multi-source data analysis. Driven by multi-source measurement data, including the axle box, the bogie frame and the carbody accelerations, this paper proposes a track irregularities monitoring network (TIMNet) based on deep learning methods. TIMNet uses the feature extraction capability of convolutional neural networks and the sequence mapping capability of the long short-term memory model to explore the mapping relationship between vehicle accelerations and track irregularities. The particle swarm optimization algorithm is used to optimize the network parameters, so that both the vertical and lateral track irregularities can be accurately identified in the time and spatial domains. The effectiveness and superiority of the proposed TIMNet is analyzed under different simulation conditions using a vehicle dynamics model. Field tests are conducted to prove the availability of the proposed TIMNet in quantitatively monitoring vertical and lateral track irregularities. Furthermore, comparative tests show that the TIMNet has a better fitting degree and timeliness in monitoring track irregularities (vertical R² of 0.91, lateral R² of 0.84 and time cost of 10 ms), compared to other classical regression. The test also proves that the TIMNet has a better anti-interference ability than other regression models.

The operational and regional conditions to which the prestressed concrete sleeper (PCS) is subjected in a railway track significantly contribute to its performance and durability. Maintaining the health of PCS poses challenges, and one of these issues involves the potential occurrence of longitudinal cracks in reinforcing bars, which can be caused by various constructional, functional, and environmental factors. Longitudinal cracks in PCS compromise the structural performance, resulting in a reduced capacity to withstand the loads exerted by moving vehicles. The current evaluations not only fail to yield a precise parameter for estimating the behavior and response of the PCS, but they also overlook the specific conditions of the PCS, such as prestressing, and only provide limited information regarding existing damage. Balancing the need for accurate evaluation with consideration of costs and resources, and making informed decisions about maintenance and track performance enhancement, has become a multifaceted challenge in ensuring a robust PCS assessment. This research introduces a novel methodology to improve the evaluation of mechanical and geometrical parameters of PCS over their operational lifespan. The objective is to enhance the accuracy of PCS performance estimation by concentrating on detecting longitudinal cracks. The suggested approach seamlessly integrates model updating methods and the finite element (FE) approach to achieve an accurate and timely assessment of PCS conditions. This comprehensive examination scrutinizes the methodology by applying artificial cracks to the PCS. In addition to introducing this assessment approach, a detailed examination is conducted on a laboratory-simulated PCS featuring various combinations of longitudinal cracks measuring 40, 80, and 120 cm in length. This systematic and rigorous approach ensures the reliability and robustness of the methodology. Ultimately, the parameters of cross-sectional area, moment of inertia, and modulus of elasticity, which significantly impact the performance of this sleeper, are explored and demonstrated through functional methodologies. The findings suggest that assessing and addressing damage should be conducted through a comprehensive and integrated procedure, taking into account the actual conditions of the PCS. Longitudinal cracks lead to a substantial decrease in the performance of these components in railway tracks. By applying the proposed methods, it is anticipated that the evaluation error for these components will be reduced by approximately 30% compared to visual inspections, particularly in predicting the extent of damage for cracks measuring up to 120 cm. This research has the potential to significantly enhance the evaluation of PCS performance and mitigate the impact of longitudinal cracks on the safety and longevity of ballasted railway tracks in desert areas.