Ultrasonic guided waves are widely used in the nondestructive testing (NDT) of aboveground pipelines. However, their application in buried pipeline inspection is significantly hindered by severe soil-induced attenuation. This study proposes a method for detecting defects in buried pipelines using nonlinear chirp signals encoded with orthogonal complementary Golay code pairs. By adjusting the proportion of low-frequency and high-frequency components in the excitation signal, the attenuation of guided waves in buried pipelines is effectively reduced. Meanwhile, the use of coded sequences increases the energy of the excitation signal, and the excellent autocorrelation properties of broadband signals enhance the time-domain resolution of defect echoes. The fundamental principles of coded excitation based on nonlinear chirp signals and pulse compression methods are first introduced. MATLAB simulations are then employed to verify the approach's effectiveness in the characterization of defect echoes under various conditions and signal-to-noise ratios (SNR). A subsequent comparative analysis, using finite element (FE) simulations for buried pipelines, demonstrates that nonlinear chirp signals with a higher proportion of low-frequency components exhibit better resistance to attenuation. By fine-tuning the chirp parameters, higher defect reflectivity can be achieved than with conventional tone bursts for various defect types in buried pipelines. FE simulations further illustrate the superiority of the proposed method over tone bursts in terms of excitation signal amplitude, defect echo reflectivity, and defect location accuracy. Finally, defect detection experiments on buried pipelines with multiple defects confirm that the nonlinear chirp signals with carefully selected parameters exhibit lower attenuation rates. In the same testing environment, the coded nonlinear chirp signals outperform tone bursts by providing higher excitation amplitudes, greater defect echo reflectivity with an increase of up to 81.45 percent, and enhanced time-domain resolution. The proposed method effectively reduces ultrasonic guided wave attenuation in buried pipelines while increasing defect echo reflectivity and extending the effective detection range.

Solar panels are essential for converting sunlight into electricity. Still, environmental factors can significantly compromise their efficiency and performance, particularly the accumulation of soiling on their surfaces or damage. This study proposes a hybrid model comprising an ensemble of deep-learning models to distinguish between soiled and damaged solar panels and their corresponding conditions. Our approach utilizes pretrained deep learning models, fine-tuned for detecting soiling or damage on photovoltaic (PV) panels, to extract relevant features and build efficient classifiers. Introducing a post-processing ensemble model improves classification metrics compared to a single deep-learning model. Combining Convolutional Neural Networks and Vision Transformers in an ensemble model achieves the highest accuracy, with 96.3% for damage and soiling detection and 91.8% for damage and soiling type identification. These results significantly outperform one-tier deep learning models, which attain an accuracy of 87.7% when classifying all possible damage and soiling categories.

Structural health monitoring (SHM) is crucial in the early stage of damage formation for the life-cycle service of offshore structures. The influence of soils on vibration-based damage detection systems in offshore structures is a critical issue but has received less attention in previous literature. Due to the complexity of offshore structures and their exposure to diverse loads, simultaneous compound damages across different components can occur, posing a significant challenge for damage detection. Existing methods often treat compound damage as a distinct type of damage, independent of corresponding single damages. Nonetheless, in cases where damages arise concurrently, the distinct characteristics of each individual damage are evident independently within the vibration signals. This study presents a new approach for detecting both single and compound damage in offshore structures considering soil interaction using vibration data. The approach combines Wavelet Transform (WT) with a Multiple Interference Deep Convolutional Neural Network (MIDCNN) to effectively learn desired features and detect damage in these structures. The MIDCNN model is trained on time-frequency data from healthy and single damage states, without incorporating time-frequency data from compound damage during training. In the testing phase, the MIDCNN model intelligently alarms healthy, single damage states, and an untrained compound damage state based on predefined probabilistic conditions derived from the MIDCNN output probabilities. The time-frequency data are generated using the WT method, which is adept at capturing the natural characteristics of the structure while minimizing the influence of noise or irrelevant components. The proposed approach is validated using measured data from a laboratory-scale offshore monopile model with soil interaction. The findings demonstrate that the proposed method is more robust than other methods in extracting features and classifying various states, including healthy, single and compound damages.

The structural soundness of a conventional track is often assessed by a single parameter called track modulus. Track modulus is a measure of the vertical deflection of the track's components beneath the rail. However, defining the track substructure's condition based only on track modulus can be misleading, as combinations of different ballast and subgrade conditions might yield the same track modulus measurement. For railroaders to be able to make an informed decision on the right maintenance strategy when a low track modulus is present, identification of the defective component between ballast or soil is critical. The railroad industry, therefore, needs an inspection technique that independently highlights the condition of the ballast and the subgrade. Addressing this challenge, our research has devised a system that helps identify the ballast and subgrade condition without disrupting normal train operations. The proposed system is a significant advancement over conventionally employed inspection methods. This new system, called the Smartgrid, uses sensors and strain gauges embedded in a geogrid sheet placed in the ballast-subgrade interface to record data on the stress-strain relationship at this plane. This data is then analyzed using supervised machine-learning techniques such as Logistic Regression and the Support Vector Machine. The ultimate objective of the proposed Smartgrid system is to arm the railroader with the right information on the condition of the two major components of the substructure and facilitate efficient maintenance. The Smartgrid, which has been tested under various conditions, promises a substantial improvement in inspection of the rail substructure.

As urban infrastructure becomes denser, subway operations face numerous risks, including adjacent construction and structural damage deterioration. Integrated monitoring is crucial to ensure safety. This study presents a multitask vibration transformer (M-TVT) model, which utilizes an Expert-Gate network to manage task-related information sharing. The model achieves integrated identification for the type and spatial distribution of internal and external tunnel abnormal vibration responses, along with assessing structural damage degree. Meanwhile, a tunnel-train-soil coupling experiment is conducted to validate the efficacy of the M-TVT model. The results demonstrate that the M-TVT model with the Expert-Gate network maintains high accuracy and stability, efficiently shares parameter information among tasks and achieves improvement in accuracy compared to single-task models.

This study examines the structural characteristics and post-earthquake damage status of 2790 buildings in Antakya downtown after the Kahramanmara & scedil; earthquakes.The study reveals how structural features, especially building typologies, construction year, number of floors, ground, and load-bearing systems, affect damage status of buildings.Damage levels resulting from the earthquake were documented, and the causes of these damages were identified.Comparisons and analyses clarify the possible vulnerability and risk status of these structures, based on studies conducted before the earthquake.This study emphasizes the importance of post-earthquake structural analyses, damage assessments, and supports future building projects and measures to be taken against earthquake risks.

Machine learning (ML) algorithms are increasingly applied to structure health monitoring (SHM) problems. However, their application to pile damage detection (PDD) is hindered by the complexity of the problem. A novel multi-sensor pile damage detection (MSPDD) method is proposed in this paper to extend the application of ML algorithms in the automatic identification of PDD. The time-series signals collected by multiple sensors during the pile integrity test are first processed by the traveling wave decomposition (TWD) theory and are then input into a hybrid one-dimensional (1D) convolutional and recurrent neural network. The hybrid neural network can achieve the automatic multi-task identification of pile damage detection based on the time series of MSPDD results. Finally, the analytical solution-based sample set is utilized to evaluate the performance of the proposed hybrid model. The outputs of the multi-task learning framework can provide a detailed description of the actual pile quality and provide strong support for the classification of pile quality as well.