Measuring the structure's displacement plays a very important role in ensuring the safe operation of railway bridges in general and urban railway bridges in particular. In Vietnam, traditional methods using high-precision mechanical gauges have been used to measure the displacement of railway bridges. However, these methods need a lot of effort in installation and traffic control during implementation. These methods are based on the static principle: The test loads are placed on the bridge structure, and then the structure's displacement is observed. The safety assessment and analysis results are guaranteed by multiplying the dynamic coefficients, leading to some assessments that may not be close to the actual exploitation of the bridge structure. Therefore, the current study presents a new solution for measuring the displacement of railway bridge structures. This method uses Linear Variable Differential Transformer (LVDT) sensors to record the continuous displacement of the structure during the time the train passes over the bridge. Through field measurements combined with a finite element analysis model, the research focuses on developing and applying LVDT sensors in urban railway bridge structure health monitoring. At the same time, the potential of developing this method in Vietnam in the future is evaluated.

Structural health monitoring (SHM) plays a crucial role in ensuring the safety of infrastructure in general, especially critical infrastructure such as bridges. SHM systems allow the real-time monitoring of structural conditions and early detection of abnormalities. This enables managers to make accurate decisions during the operation of the infrastructure. However, for various reasons, data from SHM systems may be interrupted or faulty, leading to serious consequences. This study proposes using a Convolutional Neural Network (CNN) combined with Gated Recurrent Units (GRUs) to recover lost data from accelerometer sensors in SHM systems. CNNs are adept at capturing spatial patterns in data, making them highly effective for recognizing localized features in sensor signals. At the same time, GRUs are designed to model sequential dependencies over time, making the combined architecture particularly suited for time-series data. A dataset collected from a real bridge structure will be used to validate the proposed method. Different cases of data loss are considered to demonstrate the feasibility and potential of the CNN-GRU approach. The results show that the CNN-GRU hybrid network effectively recovers data in both single-channel and multi-channel data loss scenarios.

This paper proposes a novel approach for structural damage identification using the Salp Swarm Algorithm (SSA) combined with Orthogonal Diagonalization (OD). SSA is an optimization algorithm developed in recent years and has been considered for addressing complex and computationally expensive engineering problems. However, due to the utilization of only the leader salp to search for optimal solutions, there is an imbalance between the exploitation and exploration capabilities of the salp swarm. To address this issue, we suggest the use of OD to enhance the search capability of the leader salp. OD plays a crucial role in reducing the search space and selecting high-quality salp elements in subsequent iterations. From that idea, the algorithm is tested with 12 benchmark functions, and the results demonstrate superior convergence for all of these compared to the algorithm without OD. A comprehensive study with various damage scenarios is also conducted with the Cua Rao bridge. The outcomes indicate that the accuracy, with a deviation rate when using OD, is only 1.412%, whereas it is 2.224% for the SSA in the case of a single damage scenario. As the number of damaged elements increases, the error of SSA rises to 99.667% when identifying damaged elements, while errors using OD remain below 10%. Additionally, the computational speed is significantly enhanced by approximately 40% in both scenarios.

Urban railway bridges are critical components of modern transportation networks. Dynamic loads and harsh environments put urban railway bridges at high risk of damage. Conventional vibration-based damage detection approaches often fail to provide sufficient accuracy and robustness under complex urban conditions. To address this limitation, this study introduces a hybrid deep learning framework that integrates a one-dimensional convolutional neural network (1D-CNN) and a recurrent neural network (RNN) for automatic damage detection using Linear Variable Differential Transformer (LVDT) displacement data. Start with the calibration of a finite element model (FEM) of the target bridge, achieved through updating the model parameters to align with field-acquired LVDT data, thereby establishing a robust and reliable baseline representation of the structure’s behaviour. Subsequently, a series of failure and damage scenarios is introduced within the FEM, and the associated dynamic displacement responses are generated to construct a comprehensive synthetic training dataset. These time-series responses serve as input for training a hybrid deep learning architecture, which integrates a one-dimensional convolutional neural network (1D-CNN) for automated feature extraction with a recurrent neural network (RNN) designed to capture the temporal dependencies inherent in the structural response data. Results show rapid convergence and minimal error in single-damage cases, and robust performance in multi-damage conditions on a dataset exceeding 5 million samples; the model attains a mean absolute error of ≈3.2% for damage severity and an average localisation error of <0.7 m. The findings highlight the effectiveness of combining numerical simulation with advanced data-driven approaches to provide a practical, data-efficient, and scalable solution for structural health monitoring in the urban railway context.

Giám sát sức khỏe công trình cầu bằng các biến thể nâng cấp của mạng hồi quy RNN (Recurrent Neural Network) trong xử lý dữ liệu chuỗi đã thu hút sự quan tâm lớn từ cộng đồng quốc tế. Bài báo này đề xuất sử dụng GRU (Gated Recurrent Unit), một biến thể cải tiến của RNN, để phát hiện và xác định vị trí hư hỏng trong kết cấu cầu thông qua dữ liệu chuỗi thời gian từ các cảm biến ảo. Dữ liệu này được trích xuất từ một mô hình số của cầu D2, được cập nhật dựa trên dữ liệu đo bằng cáp quang. Bộ dữ liệu được đưa vào mô hình GRU, được thiết kế đặc biệt để phân loại các nhãn hư hỏng tương ứng với dữ liệu đầu vào. Hiệu suất của mô hình GRU được so sánh với RNN truyền thống. Kết quả cho thấy cả RNN và GRU đều có khả năng làm việc trên tập dữ liệu, tuy nhiên GRU vượt trội hơn với độ chính xác trên tập kiểm tra và kiểm thử lần lượt là 81,8% và 84,3%. Điều này chứng minh rằng mô hình GRU đề xuất có khả năng chẩn đoán nhanh chóng và chính xác các trường hợp hư hỏng có thể xảy ra trong kết cấu cầu