Discrepancies between Finite Element Model (FEM) simulations and actual measurements result in substantial domain mismatches, posing a significant challenge for model-driven structural damage identification. Existing cross-domain damage identification methods commonly suffer from misalignment between domain-crossing features and damage-related features, as well as unstable network training, thereby limiting their effectiveness. To address these issues and achieve both efficient domain adaptation and high-precision damage identification, this study proposes a Frequency-Guided Cycle-Consistent Generative Adversarial Network (FG-CycleGAN), integrated with a Residual Neural Network (ResNet). First, frequency cosine similarity is introduced into the adversarial training process to quantify spectral discrepancies between generated and measured samples, ensuring the preservation of damage-relevant features during the cross-domain transformation. Subsequently, ResNet is employed to extract essential features from the samples generated by FG-CycleGAN and map them to corresponding structural damage states. To validate the approach, a damage identification experiment is conducted on a steel truss model. Comparative analysis reveals that conventional Adversarial Discriminative Domain Adaptation (ADDA) yields a relatively low F1-score of 0.35, while the basic CycleGAN achieves 0.92. In contrast, the proposed FG-CycleGAN further improves performance, attaining an F1-score of 0.99. The results confirm that FG-CycleGAN not only outperforms existing methods in terms of accuracy but also offers a robust framework for cross-domain structural damage identification.

The bonding of Carbon Fiber Reinforced Polymer (CFRP)-reinforced steel structures relies on epoxy resin, and the debonding within the adhesive layer can significantly compromise the effectiveness of the reinforcement or even lead to failure. In response to this challenge, this paper introduces a method to identify elastic constants and characterize debonding damage in CFRP-reinforced steel plates through the optimization of guided wave dispersion curves. A zero-crossing technique and spectrum decomposition were employed to extract the true dispersion curves, which simplifies the process while meets the precision requirements for elastic moduli inversion and damage characterization. The inversion errors for warping-related moduli (Gxy, Kxy) were as low as 0.63 and 0.91%, With the exception of Poisson’s ratio, the inversion errors of all elastic constants do not exceed 4%. This study represents the first application of the semi-analytical finite element (SAFE) method to the forward analysis of CFRP-epoxy-steel plates. This extension not only markedly reduces the computational cost of iterative optimization but also ensures the accuracy of parameter inversion. Both numerical simulations and experimental validations confirm that the proposed method can effectively invert material properties and evaluate Interfacial debonding damage in CFRP-reinforced steel plates, thereby demonstrating substantial potential for automated damage assessment in complex multimaterial bonded systems.

Carbon Fiber Reinforced Polymer (CFRP) structures are prone to damage under variable temperature environments. Temperature fluctuations not only accelerate damage evolution but also adversely affect guided wave-based techniques commonly used for CFRP damage detection. To address this issue, this study proposes an Autoencoder (AE) -based temperature compensation method combined with a Bayesian fusion framework using ultrasonic guided wave data. The goal is to mitigate the effects of environmental temperature fluctuations and enhance the accuracy of defect localization. The approach is proposed by training the AE with baseline signals collected under a subset of temperature conditions and then reconstructing baseline signals for other temperature by processing damage signals. Experimental validation shows that, with baseline data from only 39 temperature points, the proposed method can accurately reconstruct baseline signals at an additional 117 temperature points. Subsequently, wavelet transform is employed to extract the Time of Flight (TOF) of scattered signals, and a Bayesian data fusion framework is utilized to integrate the Reconstruction Algorithm for Probabilistic Inspection of Damage (RAPID) method with TOF information for precise defect localization. The reconstructed baselines closely align with actual measured baselines, confirming the efficacy of the proposed temperature compensation strategy. In two representative damage scenarios, the proposed Bayesian fusion method reduces the average localization errors by 38.52% and 26.43%, respectively, compared with the conventional RAPID method, while significantly suppressing imaging artifacts.

Local vibrational resonances associated with zero-group velocity (ZGV) and cutoff frequency points in dispersion curves have been extensively investigated in plate structures. These local resonances are identified as sharp and spatially localized peaks in amplitude spectra. Because local resonances can be used for nondestructive evaluation (NDE) and structural health monitoring (SHM) of waveguide structures, it is important to understand their characteristics in engineering structures. While previous studies primarily focused on plate and pipe structures, this study examines local resonances in continuous welded rail (CWR), which is widely used in modern railway systems. First, we implement semi-analytical finite element (SAFE) analysis and fully discretized finite element (FDFE) models to predict and identify ZGV and cutoff frequency points and associated local resonances in a free rail. Experimental data are then collected using surface-bonded piezoelectric patches as excitation sources in a section of free rail and CWR at a revenue-service railroad site for frequencies up to 80 kHz. The dispersion relations and local resonance spectra from numerical simulations demonstrate reasonable agreement with experimental results from both specimens, and the detectability and excitability of local resonances are discussed. Additional experimental tests reveal important energy trapping and localized minimum frequency behaviors for ZGV and cutoff frequency resonances in CWR, and furthermore the influences of rail support conditions on local resonance behavior are studied.

Damage localization methods based on Acoustic Emission (AE) can be classified into time-based and waveform- based. However, the former requires a large number of sensors while the latter is limited to 2D plane localization. In order to address the challenge of achieving more accurate 3D localization using a reduced number of sensors, this paper proposes a Circular Phased Array using Minimum Variance Distortionless Response (MVDR) Beam- forming with Autocorrelation Matrix Diagonal Loading (AMDL) method. Firstly, a sparse circular array is utilized to form multiple beamforming for coherent shear wave signals, decomposing the original 3D localization problem into Direction Of Arrival (DOA) estimation. Secondly, azimuth angle, elevation angle and autocorrelation matrix diagonal loading methods are introduced, working in conjunction with the MVDR beamforming algorithm. Finally, spatial integration is performed through matrix decomposition to solve geometric over- determined equations. The effectiveness of the proposed method is validated through numerical simulations and experimental verifications under various damage conditions. Results indicate that estimation errors for azimuth and elevation angles are both less than 2 %, while 3D damage source localization errors remain within a range of less than 3 %. This proposed method extends beamforming technology from 2D plane localization to 3D localization, significantly reducing the complexity of sensor arrangement and lowering the cost of structural health monitoring systems by utilizing a small number of sensors.

The application of Carbon Fiber Reinforced Polymer (CFRP) in reinforcing steel structures is widely recognized. However, there is relatively little research on the localization and imaging of debonding damage in CFRP-reinforced steel structures. This paper proposes a probabilistic imaging method improved by ultrasonic guided-wave transfer function to localize debonding damage in CFRP-reinforced steel structures. Firstly, this study proposes a waveform feature index that exhibits strong robustness against debonding damages while exhibiting minimal susceptibility to environmental disturbances, which enhances the detection capability for small-scale debonding damages compared to traditional linear indices. Secondly, the proposed method replaces the conventional fixed array with a dynamic scanning approach. This method achieves 2D debonding damage imaging by leveraging information solely from orthogonal directions, which not only drastically reduces the number of sensors but also enables flexible adjustment of the detection area, thereby enhancing its applicability. Thirdly, the proposed waveform feature index is independent of the amplitude of the excitation/receiving signal. Therefore, the proposed method maintains accurate localization of debonding damage during damage imaging detection, regardless of variations in coupling conditions between the sensor and the structure under inspection. The efficacy of the proposed method is validated through comprehensive numerical simulations and experiments. The results demonstrate its ability to accurately detect and localize damage in CFRP-reinforced steel plate structures, offering an effective and precise way for early debonding detection.

Noise is inevitably present in bridge health monitoring data owing to complex operational conditions, which significantly disturbs structural assessments. This study proposes a Long Short- Term Memory (LSTM)-based denoising technique to filter out noise from bridge acceleration data. Initially, the characteristics of typical noise and structural responses in bridge accelerations are investigated. Subsequently, numerical simulation provides the dataset required for LSTM training. A LSTM model is designed to take contaminated structural responses as input and output the noise component. After training, the LSTM effectively identifies and removes the noise component, enhancing the quality of the data. Moreover, acceleration data from one finite element model bridge and three in-situ bridges are used to validate this method. Both the contaminated and cleaned data are analyzed using covariance-driven stochastic subspace identification (SSI-COV) and ensemble empirical modal decomposition (EEMD) methods to assess the denoising effectiveness by stability graphs, correlation coefficients and orthogonal exponents. Furthermore, it also demonstrated that the cleaned data provides more precise modal parameter identification in both time and frequency domains. The proposed method offers an adaptive approach for noise reduction in bridge health monitoring, improving the accuracy of structural assessments.

Rail breakage can cause train derailment that leads to catastrophic consequences, prevention of which is worthy of a significant amount of investment in finance and time. Non-destructive evaluation can provide qualitative and quantitative inspection of rail tracks in-situ to detect dangerous faults, even though these defects are primarily invisible or hardly noticed from the surface of the rail. To make it efficient and convenient, infrared thermography technique offers a remote operation without interfering with rail transportation. The image processing for edge detection can further quantify the defect deterioration by depicting the contour of the defect. Moreover, the optimal parameters for various edge operators provide reliable detection for natural fracture defects.

With the continuous expansion of waterborne transportation, the risk of collisions between ships and lock gates has gradually increased. Therefore, it is urgent to clarify the dynamic response patterns of lock gates during impacts and to assess their risk consequences. This study employs the Abaqus/Explicit solver to analyze the dynamic response patterns of miter gates under ship impacts, aiming to identify the most hazardous impact scenarios and vulnerable parts. The impact of ship stiffness and the added mass of fluid on the collision has been considered. Initially, tension tests on Q390D specimens were conducted to derive the constitutive model. The accuracy of the simulation was verified based on the analytical solution of impact force and the principle of energy conservation. Subsequently, by analyzing the dynamic response, vulnerability, and crashworthiness of miter gates, the study further explored the mechanisms of influence regarding ship impact tonnage, speed, position, and angle. The results indicate that ship speed has the most significant effect on gate damage, which should be restricted to within 1 m/s. The central area of the miter gates experiences more severe deformation from impacts than the side areas. Finally, models for the impact force of ships and the deformation of miter gates are introduced, enabling rapid and efficient of impact forces and gate deformation under various scenarios. The findings provide strong support for the reinforcement design and safe operation of miter gates.

Low yield point (LYP) structural steels had been developed and used for the seismic energy dissipating, however, the fracture of various LYP steel energy dissipation components may lead to progressive collapse of structures. To address this issue, the fracture behavior of LYP steels under various stress states were comprehensively investigated using experiments, theoretical analysis, and numerical simulations in this work. There are three meaningful contribution points. First, an improved weighted average (WA) method and dichotomy-based optimization are proposed to calibrate true stress-strain relation after necking onset. The maximum deviation for the prediction of full-range tensile load-displacement behavior based on the proposed WA method is less than 3%, which is better than the existing WA methods. Second, a modified ductile fracture model is developed to identify the fracture of LYP steels under various stress states. Comparison results among existing fracture models, the proposed fracture model and tests indicate that the fracture plastic strain can be more reasonably predicted by the proposed ductile fracture model than the existing models. Thirdly, an advanced ductile fracture prediction framework is proposed to simulate the complete tensile behavior of structural steels. Based on the proposed framework and object-oriented programming technique, the complete tensile behavior can be accurately and conveniently predicted by employing the proposed WA model, proposed ductile fracture model and developed user defined material subroutine. The verified WA method and ductile fracture model would contribute to predicting the failure of structural steel members, connections, welds and structures. It is meaningful to prevent casualties and property losses resulting from the subsequent structural progressive collapse.

Thermographic techniques are widely used to characterise material damage, porosity and moisture in structures. Due to the thermographic signal reconstruction (TSR) technique, active thermography has made significant progress. In this study, the TSR technique was applied to identify a rubber layer that mimics an internal defect, which is attached to the back of the steel plate, after a 1.0 min thermal stimulation. The peaks of the first and second derivative logarithmic curves correspond to the important time point revealing the position of the attached flaw. Furthermore, a new post-processing technique for infrared thermographic data was proposed by exploiting polynomial orders of the fitting curves to produce a synthetic image, which can highlight the flaw with distinctive contrast. Finally, the upgraded TSR technique was applied to image a natural shelling defect in a rail sample with an excellent contrast.

Time–frequency decomposition is a powerful tool in assessing the dynamic behaviors of structures. Traditional time–frequency decomposition methods struggle with adaptability, and are limited in handling the structural responses with strong nonlinearity and closely spaced modes. In this study, a cutting‐edge approach based on deep neural network (DNN) is proposed to achieve a precise and adaptive time–frequency decomposition of nonlinear structural responses under seismic excitations. Remarkably, despite being trained on synthetic samples, the proposed method demonstrates outstanding performance in decomposing time–frequency components across various seismic response scenarios. Compared to variational mode decomposition (VMD) and synchroextracting transform (SET), the proposed method exhibited superior precision in time–frequency decomposition and excellent efficiency in parameter optimization. Moreover, the applicability of the proposed method to real‐world complex structures has been verified, which also shows promising generalization capabilities. Future research will aim to enhance the network performance by incorporating additional learning samples with diverse nonlinear characteristics.

This study showcases the electromechanical impedance (EMI) technique for extracting and promoting zero-group velocity (ZGV) and cutoff frequency resonances in a waveguide structure. We identify the mechanisms of multiple resonances in the EMI spectra via a wave propagation perspective. Both simulation and experiments reveal the fact that sharp resonances in the conductance spectra are associated with either ZGV or cutoff frequency points. Consequently, we design four test configurations to enhance local resonances by aligning induced motions with considered mode shapes. Reasonable agreement between simulation and experiment results is observed. We evaluate the performance of considered configurations in terms of mode enhancement, and configurations that can selectively promote certain mode families are summarized. This study also shines the light on the EMI technique for quantitative non-destructive evaluation (NDE) by potentially supporting the inverse characterization of mechanical properties of host structures.

Damage localization is one of the most challenging topics within Structural Health Monitoring (SHM) in aeronautics, especially when the structure is manufactured out of carbon fiber-reinforced composite materials. Using ultrasonic guided waves (particularly Lamb waves), generated and recorded with piezoelectric transducers, is also challenging in this type of material. Otherwise, traditional methods used for this task are subjected to physics-based knowledge of the problem, such as damage imaging algorithms like delay-and-sum and RAPID. This paper presents an entirely data-driven approach, based on the ability of Deep Learning (DL) techniques (particularly those based on Convolutional Neural Networks – CNNs –) to extract features of interest for damage imaging from a pre-dataset. In this work, the selected feature to be estimated is the normal distance from the propagation path of the guided wave to a simulated damage, which allows, in combination with an especially designed positioning algorithm, to locate with high accuracy defects, even in different positions than the used for the training of the network (a fixed grid of points over the analysis zone). This paper presents the application of the method to a real composite material specimen, as well as the recorded results obtained from additional datasets recorded with the simulated damage (a piece of blu-tack) attached to different random positions other than those of the training grid.

Accurate and reliable nondestructive evaluation (NDE) for stress measurement is essential for assessing structural performance and preventive maintenance. Local resonances offer an efficient NDE method due to their "amplified and localized" vibration amplitude. This allows for mounting or wiring near the structural edge without affecting the vibration modes. In this paper, we investigate the influence of stress level on the local resonances in a rectangular aluminum bar structure. In the first step, we utilize the electromechanical impedance method (EMI) to extract local resonances and verified them as Zero-Group-Velocity (ZGV) mode and cutoff frequency modes by performing two-dimensional fast Fourier transform (2D-FFT) on the wavefield. Furthermore, we investigate the influence of uniaxial tensile stress on the local resonance frequencies. It is observed that both the ZGV mode and cutoff frequency mode demonstrate a measurable sensitivity to the applied axial load.

An ultrasonic sonar-based ranging technique is introduced for measuring full-field railroad crosstie (sleeper) deflections. Tie deflection measurements have numerous applications, such as detecting degrading ballast support conditions and evaluating sleeper or track stiffness. The proposed technique utilizes an array of air-coupled ultrasonic transducers oriented parallel to the tie, capable of “in-motion” contactless inspections. The transducers are used in pulse-echo mode, and the distance between the transducer and the tie surface is computed by tracking the time-of-flight of the reflected waveforms from the tie surface. An adaptive, reference-based cross-correlation operation is used to compute the relative tie deflections. Multiple measurements along the width of the tie allow the measurement of twisting deformations and longitudinal deflections (3D deflections). Computer vision-based image classification techniques are also utilized for demarcating tie boundaries and tracking the spatial location of measurements along the direction of train movement. Results from field tests, conducted at walking speed at a BNSF train yard in San Diego, CA, with a loaded train car are presented. The tie deflection accuracy and repeatability analyses indicate the potential of the technique to extract full-field tie deflections in a non-contact manner. Further developments are needed to enable measurements at higher speeds.

Local resonances, formed by zero-group velocity (ZGV) and cutoff frequency points, have been extensively studied using impulse-based approaches, such as pulse laser and impact echo. In this work, we showcase the electromechanical impedance (EMI) technique as an option to extract and promote zero-group velocity and cutoff frequency resonances in a waveguide structure. We identify the mechanisms of multiple resonances in the EMI spectra via a wave propagation perspective. Furthermore, we extract the dynamic response profiles at a cutoff frequency and a ZGV frequency to confirm the localized minimum frequency behavior within corresponding branches.

This paper presents an ultrasonic technique for non-contact surface deflection mapping of railroad ties using concepts of sonar-based ranging. The technique utilizes an array of capacitive ultrasonic transducers arranged along the length of the railroad tie at a lift-off distance of 3 in. from the rail surface to ensure contactless measurements. The transducer array is used in pulse-echo mode and distances from the transducers to the tie surface are measured by tracking the time-of-flight of the waves reflected from the tie surface. A reference-based cross-correlation operation is introduced to compute the time-of-flight, wherein one of the transducers is used as a reference for the distance measurements. The reference-based cross-correlation ensures accurate peak -detection for time-of-flight based differential distance measurements. An acoustic signal strength-based tech-nique is utilized to differentiate between signals reflected from ties and those reflected from ballast. Laboratory scale tests were first performed as a proof-of-concept on a slender wooden beam to measure deflections in loaded and unloaded conditions. Dynamic tests were also performed on this beam to determine the ability to track time -varying positions. Field tests on a replica test track with wood ties were performed at the Rail Defect Testing Facility at University of California San Diego by mounting the prototype on a test-cart moved at walking speed. A dynamic assessment of the prototype was also performed to determine natural frequencies of the mounting beam that may be relevant for future uses at higher speeds. The results indicate the potential of this non-contact system to measure full-field tie deflections in 3D in-motion. This ability may potentially be used to detect conditions of poor tie support that may cause derailments.

Local resonances formed by zero-group velocity (ZGV) and cutoff frequency points usually demonstrate sharp resonance peaks in frequency spectra, which can be utilized for nondestructive evaluation (NDE) and Structural Health Monitoring (SHM). The existence and application of those local resonances have been extensively reported in plate and pipe structures. However, local resonances in rails are rarely studied. The team recently reported that impulse dynamic tests can promote the local resonances in rails up to 40 kHz, and the results were verified using both semi-analytical finite element (SAFE) analysis and frequency-domain fully discretized finite element analysis. In this work, we present the discovery of ZGV modes and cutoff frequency resonances in free rails up to 80 kHz using piezoelectric elements. A miniature low-cost PZT patch works as a consistent excitation source compared with the impulse dynamic testing method. First, we implement the SAFE analysis to compute dispersion curves of a standard AREMA 115RE rail and to identify potential ZGV and cutoff frequency points up to 80 kHz. Then, to understand the existence and detectability of identified ZGV and cutoff points in a free rail, we install one PZT patch on the side of the rail head. A chirp signal covering 20 to 120 kHz is selected as the excitation to cover the desired frequency range. Finally, we perform a spatial sampling of wave propagation using three receivers along the wave propagation direction to calculate the dispersion relations experimentally via two-dimensional Fourier Transforms (2D-FFT). This study verifies the existence of ZGV modes in free rail up to 80 kHz and demonstrates the feasibility of using piezoelectric elements to generate local resonances.

The detection and localization of structural damage in a stiffened skin-to-stringer composite panel typical of modern aircraft construction can be addressed by ultrasonic-guided wave transducer arrays. However, the geometrical and material complexities of this part make it quite difficult to utilize physics-based concepts of wave scattering. A data-driven deep learning (DL) approach based on the convolutional neural network (CNN) is used instead for this application. The DL technique automatically selects the most sensitive wave features based on the learned training data. In addition, the generalization abilities of the network allow for detection of damage that can be different from the training scenarios. This article describes a specific 1D-CNN algorithm that has been designed for this application, and it demonstrates its ability to image damage in key regions of the stiffened composite test panel, particularly the skin region, the stringer’s flange region, and the stringer’s cap region. Covering the stringer’s regions from guided wave transducers located solely on the skin is a particularly attractive feature of the proposed SHM approach for this kind of complex structure.