This PDF file contains the front matter associated with SPIE Proceedings Volume 12487, including the Title Page, Copyright information, Table of Contents and Conference Committee lists.

To reduce CO₂-emissions lightweight structures needs to be implemented in all transport applications. At the same time, low-weight and high performance materials must provide safety and reliability, at economical prices. Extended Non-Destructive Testing (ENDT) contributes to safeguarding the performance of adhesively joined load-critical structures, permitting to steadily monitor adherent surfaces prior to bonding and to detect adhesion properties of bonded components. In the present work, approaches exceeding the state-of-the-art of innovative ENDT techniques like robot-based Laser-Induced Breakdown Spectroscopy (LIBS) are presented. Furthermore, automated, AI-based image processing and evaluation methods for surface quality inspection are shown, aiming at overcoming today’s limitations concerning handling, evaluation speed and reliability of results. First results of automated in-line surface quality assurance approaches for assessing multi material adherent surfaces are highlighted.

Subsurface inspection of concrete structures using electromagnetic (EM) sensors such as ground penetrating radar (GPR) and synthetic aperture radar (SAR) is a field applicable approach for critical civil infrastructure systems. Compared to other nondestructive inspection/evaluation/testing techniques, EM waves can penetrate the surface of concrete structures and travel inside the subsurface of concrete structures to generate backscattering signals from a subsurface target (e.g., corroded steel rebar, delamination/cracking) for condition assessment. However, variations in the EM property of concrete and unpredictable EM background noises can contribute to the difficulties of image interpretation. Denoising of radar images is a necessary step before engineers can perform quantitative assessment of the images. The objective of this paper is to denoise GPR images using discrete wavelet transform (DWT). Four concrete panel specimens (30-by-30-by-3.5 cm3 ) were prepared with three artificial cracks (CNC, CNCD, and CNCW) of known dimensions and subjected to B-scan inspection using a 1.6 GHz GPR sensor (StructureScan Mini, GSSI). Level five Daubechies wavelet was used in processing all GPR B-scan images for its capability of detecting high frequency components in this study. The purpose of image denoising was to reveal a clear hyperbolic pattern by eliminating undesired local maximum and minimum points. After each image processing, horizontal, vertical, and diagonal details were generated. Four different denoising schemes were considered; i) without all details, ii) horizontal detail only, iii) vertical detail only, and iv) diagonal detail only. Denoising was repeated in five steps in each scheme. Performance of denoising was evaluated by the number of local maximum and minimum points and their geometric pattern. From our results, it was found that, for some type of artificial crack (CNCD), the hyperbolic pattern can be clearly revealed after one step of denoising, regardless of the denoising scheme. Among four denoising schemes, the best scheme is the one with the approximation coefficients and the diagonal detail coefficients. It was also found that including horizontal detail can introduce high frequency artifacts, resulting in an over-denoised image.

Structural Health Monitoring (SHM) using Ultrasonic Guided Waves (UGWs) offers great potential in detection of minor flaws, due to the employed short wavelengths. A bottleneck in UGWs-based schemes lies in the extensive computational costs for evaluating the associated wave propagation models. Such detailed models form though a necessity to reach higher levels of SHM, e.g. localization and assessment of flaws. Reduced Order Models (ROMs) and surrogate models allow for lowering the substantial numerical costs for SHM applications, especially if they are parameterized with respect to the characteristics of different flaw configurations. Machine Learning (ML) algorithms can be trained for this purpose, however, in the case of black box ML algorithms, this comes with the drawback of the requirement for substantial data availability for the purpose of training. Such, training data, which are typically derived from full order numerical simulations, are computationally costly to obtain. To reduce the amount of training data, known information on the mechanical behavior can be harnessed and inserted into the estimation process. In the present work, a method is introduced that exploits the properties of the interaction of UGWs with flaws in the frequency domain. It can be shown that the frequency domain response is characterized by periodic features that are linked to the flaw location. An ML model based on this knowledge can be trained with less training data. The potential of this approach for damage localization in the context of SHM is illustrated in a simulated example of a composite plate.

Lithium metal batteries are prone to subtle defects such as internal dendrites, which can cause internal short circuits and lead to catastrophic ignition. These defects are often undetectable by battery management systems, prompting the need to advance the development of nondestructive evaluation (NDE) techniques for battery applications. Ultrasonic inspection techniques are being evaluated as a means of identifying flaws and irregular lithium plating that can be a precursor to dendrite formation and, ultimately, battery failure. Two ultrasonic approaches were compared in this study to assess their relative merits for battery inspection. The first was local ultrasonic resonance spectroscopy (LURS), which measures the local through-thickness resonances of the battery to detect changes in structure. The second technique was guided wave ultrasound, which was assessed for its potential for in situ monitoring. Guided wave testing was performed via pitch-catch testing using piezoelectric electric wafer active sensors (PWAS), as well as line scans via laser Doppler vibrometry (LDV). Both measurement modes were applied to lithium metal pouch cell batteries seeded with lithium chips emulating localized plating. The results show the ability to detect and monitor the internal structure of batteries for relatively coarse defects and highlight use cases for each of the two inspection modalities.

Artificial intelligence (AI) and virtual/augmented reality (VR/AR) are facilitating objective and fast assessment of infrastructures. Computer vision advancements are also transforming traditional methods into automated information modeling and decision support systems. These advancements offer new opportunities to combat growing challenges that threaten infrastructure systems. In particular, climate change, aging structures, and population growth have intensified threats to infrastructure, requiring methods for evaluating infrastructure quickly after a disaster. VR uses cameras and sensors to provide images of the current state of a structural system, including the pattern of concrete cracks. A novel framework for analyzing the degree of damage in cracked reinforced concrete shear walls (RCSWs) is presented in this paper by leveraging virtual reality (VR) technology. An automated and unbiased approach is enabled by converting images of crack patterns on the surface of concrete structures into graphs. It is possible to extract relevant information from graphs, including graph features, to quantify the extent of damage using graph theory. A machine learning algorithm is then trained using these features to predict the extent of the damage. To validate the approach, the framework was applied to data collected from three RCSWs that were subjected to quasi-static cyclic loading. ML was used to predict the secant stiffness and its descending trend during each load cycle in the experimental test. The VR-based approach achieves high R² scores of 0.90, 0.91, and 0.99 in a machine learning regression, indicating the framework's success.

Synthetic aperture focusing technique (SAFT) is one of the most-used ultrasonic imaging algorithms. However, by considering only the direct sound path (i.e., pulse-echo signals) for image reconstruction, this technique is unable to show steeply inclined interfaces and bottom boundaries of objects in concrete structures. To address this problem, the current study considered various sound paths when applying SAFT for concrete elements. The proposed hypothesis is that different sound paths will provide access and imaging opportunities to previously inaccessible areas of the inspected volume/region. As a proof of concept, the proposed method, namely the ray-based SAFT, was tested with ultrasonic shear wave data for two concrete slab specimens. The results indicated a significant improvement compared to conventional SAFT imaging for simulated rebar debonding, vertical boundaries of slabs, and bottoms/sides of tendon ducts.

We demonstrate the enhancement of contact nonlinear ultrasonic testing by implementing the 1-D phononic crystal layers. Nonlinear ultrasonic testing (NLUT) can detect sub-wavelength defects in concrete using the second harmonic generation method (SHG). SHG method is based on the generation of the second harmonic signal in solids when the first harmonic wave interacts with defects. Due to the inherent heterogeneous characteristic of concrete, the typical ultrasonic frequency for NLUT in concrete is 50 - 100 kHz. Frequency higher than 100 kHz may introduce high nonlinearity due to the multiple scattering in concrete, which would mask defect-related nonlinearity to be detected by NLUT. A 1-D phononic crystal is placed between the sensor and concrete surface to enhance the signal quality to block the signal beyond 100 kHz exhibited from instrumentation and couplant. The 1-D phononic crystal structure consists of a periodical arrangement of composite layers that can tune the band gap 100 kHz-200 kHz. The binary composite system configuration using the mixture of steel and aluminum is numerically studied. The periodic array exhibiting a band gap near 100kHz-200kHz is selected for experimental validation. The sensitivity of NLUT is improved by detecting the sub-wavelength inclusion in concrete. The results demonstrated that the nonlinearity due to defects becomes more apparent by removing the instrumentation-induced nonlinearity in concrete.

Half grouted sleeve connection has been widely used in the rebars connection of prefabricated concrete (PC) structure. Mostly, the implementation of grouted should be finished on site. Meanwhile, the internal defects are inevitable due to the concrete nature. Currently, there is few methods available, which can effectively and rapidly evaluate the quality of the connection. Therefore, in this paper, we propose a combination of low-frequency linear ultrasound (LUT) and nonlinear ultrasound (NLUT) to quantitatively characterize defect. The internal artificial defects are concentrated defects, and the defect content is 10%, 20%, 30% and 40% respectively. Through transmission mode was adapted for both LUT and NLUT. The UT wave propagation was distorted by different defects, which was the results of LUT. For NLUT with higher resolution, the complex distribution and different level of defect together will introduce nonlinearity. The experimental results show that the grouted defects reduce the ultrasonic energy of LUT, and increase the nonlinearity from NLUT with the increase of the defect size and randomness. The defect has a significant impact on the ultrasonic features. Therefore, Low-frequency LUT and NLUT methods are potential to realize the visualization the defects of half grouted sleeve connection.

Fiber-reinforced polymer matrix composites deteriorate mechanically due to fatigue degradation during cyclic stress. The progressive decrease in elastic stiffness over fatigue life is well-established and investigated, yet many dynamic engineering systems that use composite materials are subjected to random and unexpected loading circumstances, making it impossible to continually monitor such structural changes. LIG can detect strain and damage in fiberglass composites under quasi-static and dynamic loads. ANNs and traditional phenomenological models may assess damage development and fatigue life utilizing LIG interlayered fiberglass composites’ piezoresistivity. Passive experiments monitor LIG interlayered fiberglass composite elastic stiffness and electrical resistance during tension–tension fatigue stress. Electrical resistance-based damage metrics follow similar trends to elastic stiffness-based parameters and may accurately depict damage development in LIG interlayered fiberglass composites over fatigue life. In specimen-to-specimen and cycle-to-cycle schemes, trained ANNs and phenomenological degradation and accumulation models predict fiberglass composite fatigue life and damage state. In a specimen-to-specimen scheme, a two-layer Bayesian regularized ANN with 40 neurons per layer beats phenomenological degradation models by at least 60%, with R² values more than 0.98 and RMSE values less than 10⁻³ . A two-layer Bayesian regularized ANN with 25 neurons per layer exhibits R² values more than 0.99 and RMSE values less than 2×10⁻⁴ when more than 30% of the original data is used in a cycle-to-cycle method. Piezoresistive LIG interlayers and ANNs can correctly and constantly predict fatigue life in multifunctional composite structures.

High level nuclear spent fuel canisters have been designed and used for nuclear waste storage since the 1950’s. Stress corrosion cracking (SCC) has shown to occur under certain corrosive chemical conditions when the residual stress was not relived in a welded plate. Typical SCC would eventually cause catastrophic failure of structures. In the case of spent nuclear canisters, the radioactive materials may leak through the cracks if they penetrate the tank wall. Early detection of SCC is crucial, followed with appropriate mitigation methods. Various mitigation methods have been funded and explored for the nuclear facilities. Among them, engineered composite patch repairing technique that was originally developed and adopted for aerospace aluminum structures has been proposed as one of the solutions. The technique begins with describing the sample preparation procedures. Based on nondestructive evaluation (NDE) techniques, the bonding between the composite patch and the repaired steel plate was then thoroughly examined using ultrasonic Lamb wave modes. The results were analyzed, and our findings demonstrated that the bonded composite patch was an effective way for crack mitigation. It was observed that they reduced the wave interactions and modified the Lamb wave modes propagation. The results were also helpful in determining the differences between the fully bonded sample and the flawed bonded sample.

Structural adhesive bonding is a growing trend in the aerospace industry. The main benefits of using such technique are elimination of drilled holes and thus micro-cracks and stress concentrations, reducing weight, and having faster assemblies. However, it is still not extensively used in critical components of various structures as inspections may become exhaustive. This is why further structural health monitoring techniques are being developed to monitor adhesive bonded joints. In this work, we monitor, in real-time, the curing cycle of adhesives used for structural bonding of carbon fiber reinforced polymers (CFRP) via guided Lamb waves. The in-situ monitoring is done experimentally inside an oven using piezoelectric transducers on a bonded structure composed of two pre-cured woven CFRP plates adhered together using structural adhesive film. The degree of cure and other cure parameters such as gelation and vitrification of the adhesive are extracted experimentally from the velocity and voltage curves. Then, using a computational finite element model in COMSOL, we further investigate the monitoring results by combining solid mechanics and electrostatics modules, and actuating a single anti-symmetric mode. The computational cure monitoring process of the adhesive is built by importing dynamic mechanical analysis (DMA) results into the numerical model. Furthermore, co-cure monitoring of both uncured CFRP and adhesive film is studied via a reusable flexible PTFE sensing film that was previously designed. Results of the latter experiment show that the A0 mode amplitude is more sensitive towards the epoxy cure parameters.

Registration techniques play a central role in applications of image processing to computer vision, medical imaging, and automatic target tracking. Feature-based techniques such as scale-invariant feature transform (SIFT) and speeded up robust features (SURF) are commonly used to register images derived from a single modality. However, SIFT and SURF struggle to register images from different modalities because the features tend to manifest rather differently and at sometimes very different length-scales. The most successful methods that have been developed to register multi-modal data use information-theoretic approaches. These methods play a key part in nondestructive evaluation scenarios where data that is collected by sensors of different modalities must be registered to be fused. In this paper, automated registration based on normalized mutual information is applied to align data derived from ultrasonic and radiographic inspections of (i) additively manufactured titanium alloy test coupons, and (ii) thin, lithium metal pouch-cell batteries. The quality of the registration is quantified in terms of computational resources and spatial accuracy. In the first case the X-ray computed tomography (XCT) data is captured on a region corresponding to a small subset of the ultrasonic data, while in the case of the lithium batteries the digital radiography (DR) captures a larger region of interest than the ultrasonic data. In both cases the radiographic data resolution is much higher than for ultrasound, but interestingly, in both cases the accuracy of the registration is approximately equal to two-to-three-pixel lengths in the ultrasonic images.

The strength of concrete near the surface would be greatly reduced subjected to high temperature, usually accompanied by surface cracking and spalling. In this study, two concrete slabs heated to 600°C and 800°C on one surface were tested using a small hammer and a displacement receiver placed on the two ends of a test line. The dispersion curve was obtained by performing Short-Time Fourier Transform and amplitude reassignment technique on a received displacement waveform. Combining 10 dispersion curves obtained from multiple grid lines, a three-dimensional surface wave velocity contour map is constructed. Concrete cores were taken after NDT testing and compared with the contour map of the location where the core was taken. The test results show that both the concrete strength and the crack depth affect the surface wave velocity. This technique allows rapid assessment of the cross-sectional wave velocity distribution of a surface, even on rough, spalled surfaces.

Ice formation and accretion on windows of buildings and windshields of automobile lead to various inconveniences and operational difficulties in cold regions. To prevent ice accretion and fasten ice removal on glass, a new generation of transparent deicing materials with high efficiency and energy saving is highly expected through low-cost approaches. However, conventional anti-icing/deicing coatings are opaque materials that are difficult to implement on automotive glass and building windows. As a typical photothermal semiconductor material, Cu2-xS has high near infrared (NIR) light absorbance and excellent photothermal conversion realized by excitation and relaxation of electron-hole pairs, which differs from noble metal nanoparticles. The unique advantage makes Cu2-xS is used widely in photothermal tumor and cancer therapy, while the application in anti/deicing area is limited. Here, we develop a low-cost transparent photothermal nanocomposite coating based on solution-processed Cu2-xS for active photothermal deicing. The photothermal nanocomposite coating was first prepared by the integration of Cu2-xS nanoparticles and commercially available acrylic paints, and then brushed onto glass surfaces of automobile and buildings. The deicing results show that when exposed to the near-infrared laser illumination at the wavelength of 808 nm, the surface coating temperature of glass covered with 3mm ice layer rapidly increases over 30℃ and the ice layer melts in 300 seconds at different ambient temperatures of -16 ℃, -20℃ and -24℃, demonstrating the high light-to-heat conversion efficiency and remarkable deicing property of transparent photothermal coating under extreme cold conditions. This study of transparent photothermal nanocomposite coating fabricated by simple brushing method provides enormous potential for ice removal applications on glass in building structures and automobile without highly affecting the visible transmittance, which is expected for further development in various shaped components.

Uniform distribution and strong interaction of carbon nanotubes in a polymer matrix have resulted in improved performance of the pure polymer. In this work, MWCNTs (CNTs) and carboxylic acid functionalized MWCNTs (CNTs-COOH) were dispersed into the epoxy resin to improve the intrinsic damage sensing property of woven carbon fiber reinforced epoxy (CFRP) composite laminates. We prepared the MWCNT-modified CFRP composites (CNT_CFRP and CNT-COOH_CFRP) with varying MWCNT concentrations (0.25 wt% ˗ 1 wt%) and studied the effect of functionalization and concentration of MWCNTs on the mechanical, and electrical properties of CFRP laminates. A pronounced increase in the tensile strength and tensile modulus of CFRP laminate was achieved with the introduction of 0.5 wt% CNTs and CNTs-COOH. The damage sensing capability of modified CFRP laminate was investigated by measuring electrical resistance upon loading.

The objective of this research is to demonstrate the versatility of a dip coating process for the efficient integration of piezoelectric barium titanate (BaTiO₃) microparticles on a wide variety of fibers to design passive self-sensing composites. The microparticles were deposited on glass, aramid, and basalt fiber weaves through the proposed dip coating technique. A computational framework is established to predict the deposition thickness on the fiber surfaces from the given microparticle concentration, size, coating velocity, and coating fluid viscosity. The deposition quality assessment was performed through scanning electron microscope imaging and subsequent image analysis. BaTiO₃-coated fibers were directly used in composite preparation. After fabrication, the BaTiO₃-enhanced composites were subjected to high-voltage poling. Finally, their passive self-sensing properties were characterized through experimental studies. The results show the adaptability of the proposed coating process to integrate BaTiO₃ microparticles within different types of fiber-reinforced composites enabling passive self-sensing to attain subsurface damage characterization.

Non-destructive evaluation (NDE) methods, like ultrasonic inspection, face numerous challenges detecting defects within thin laminates. Short backwall distances paired with wave scattering and high attenuation caused by innate properties of stacked laminates result in difficulties with distinguishing sub-surface defects, such as delamination from geometric features. To improve the effectiveness of ultrasonic inspections for thin composite laminates the authors have implemented a multi-step approach to predict wave propagation behavior. This plan focuses on the use of finite element methods to simulate wave propagation behavior utilizing a single element transducer: (i) in water; (ii) in isotropic materials; (iii) in thin composite laminates. Analytical verification based on the Fresnel approximation was performed for ultrasonic wave propagation in water prior to modeling scenarios with isotropic materials and thin composites. In addition, experimental validation was performed in parallel to gauge the accuracy of all simulations. Initial computational results using a single element transducer for on-axis pressure measurements showed that acoustic elements were well suited for modelling water paths due to their computational efficiency and accuracy when paired with continuum stress elements. Preliminary experimental results showed it was possible to physically detect defects in both isotropic and composite material cases. The authors believe that computational models can be used for evaluating complex ultrasonic signals, aiding in the development of new ultrasonic inspection techniques for thin composite structures.

This paper proposes an ultrasonic scanning system for the inspection of stiffened composite panels used in modern aircraft construction. Conventional ultrasonic scanning systems usually track individual features of the propagating waves (e.g. amplitude, velocity, etc. at specific frequencies). The proposed scanning system, instead, is able to track the full dispersion curve across a significant frequency range, and, through an inversion process, to identify the elastic constants of the panel at each scanning position for an enhanced inspection. The proposed scanning system utilized a “single-input-dual-output" (SIDO) scheme whereby ultrasonic guided waves are excited by an impact and detected by two air-coupled ultrasonic sensors. At each scanning point, the system extracts the phase velocity dispersion curves of the panel via a phase spectral analysis of the measured waveforms. The measured dispersion curve is then fed to an inversion algorithm that identifies the composite engineering constants through an optimization loop. The Semi-Analytical Finite Element (SAFE) method is used as the forward model in the inversion procedure. Validation experiments were performed on a stiffened skin-to-stringer CFRP panel with impact-type damage present on the hat shape stringer cap. The system showed an ability to detect the internal damage with access only to the outer skin. The ability to track the elastic constants of the test part is quite relevant to the ultimate goal of determining the part residual strength using the ultrasound scanning system.

This paper presents an improved damage detection technique for composite structures using a steady state excitation based on the use of a laser doppler vibrometer (LDV). Employing a single, fixed frequency excitation from a mounted piezoelectric transducer, the full steady-state wavefield could be obtained using a LDV with a mirror-tilting device. After scanning completed in high speed, a wavenumber filtering is applied to determine the dominant wavenumber components of the measured wavefield, which could be used for a damage sensitive feature. In this process, the directional correction of the wavenumber is performed to minimize the anisotropic characteristic in local wavenumber estimation. For validation of this proposed technique, several experiments were performed on composite structures with delamination damage. The results showed that the proposed technique is very efficient in detecting and quantifying delamination and debonding damage on composite structures.

In the present work, a novel combination method of in-line monitoring and offline non-destructive evaluation was developed for the detection and monitoring of defects in additively manufactured specimen. The new methodology includes Infrared Thermography, Acoustic Emission and Micro-computerised Tomography to allow for the detection of anomalies during the printing process and the verification of their presence after the printing process without the need for destructive testing. It was found that the in-line monitoring can provide information on the efficacy of the printing process which is substantiated by the offline assessment.

Acoustic Emission (AE) source localization is an effective technique to monitor the invisible damage in structures. However, when multiple damage mechanisms coincide, a single AE hit may summate multiple AE events. Concrete structures, especially 3D geometries such as slabs, may have multiple crack initiations simultaneously, which influence the arrival time accuracy. The source localization is highly affected by the selected arrival time picking method. Most commercial data acquisition systems use a threshold-based method to extract the arrival time. Due to the occurrence of multiple cracks simultaneously and the influence of system noises, signal with low signal-to-noise ratios or the continuation of a prior event in the pre-trigger regime are more likely to have more significant errors. This paper implements the conventional threshold-based method, floating threshold-based method and Akaike’s Information Criteria (AIC) method. In addition, the wavelet transformation technique was applied to the waveform to decompose the complex multi-frequency signal into a family of single-frequency wavelets. Damage-related wavelet was selected based on the wavelet coefficients in the wavelet spectrogram. This paper compares these two methods' results using raw and decomposed signals. Efficiency of the results were verified by conducting the source localization algorithm. The source localization results illustrated that the accuracy of the damage localization is significantly improved by adapting the wavelet transformation on threshold-based since the arrival time was picked more accurate.

Due to its ability to manufacture a single, complex part that either (1) could not have been built with traditional manufacturing or (2) would require the assembly of several sub-components, the use of Metal Additive Manufacturing (MAM) has increased over 300% in just the last 5 years1,2. Even with the advancements in the technology of Metal Additive Manufacturing, the final as-built mechanical properties continue to possess variability that make it difficult for designers and engineers to utilize them effectively, especially on mission critical parts. The final material properties of an MAM part are process and operator-dependent and the MAM process is more complex than traditional metal manufacturing techniques. The ability to non-destructively validate the quality of an MAM part is critical as the utilization of metal these parts increases. This paper describes ongoing research that focuses on techniques that can predict if a given part meets customer requirements and is free from hidden defects, flaws, non-visible geometric fluctuations, and variations in mechanical properties that are critical to the design/analysis process and mission certification. The techniques use Laser Doppler Vibrometry to detect the frequency response spectrum for a given part then process that data with non-p-value statistical techniques (Machine Learning) to develop models that can be “trained” to detect a variety of quality issues. Machine Learning models work best when data from many samples with different unacceptable end-states are available and used to train and develop the models. These end-states could be the result of a variety of unacceptable variations in the MAM process, quality of the precursor powdered metal, quality of the millions of micro-welds that make up the MAM build process, size and number of inclusions and micro flaws in the material as well as residual stresses in the part. In this study, our goals were to (1) use a large enough number of parts to ensure the validity of the machine learning model and (2) focus on geometric defects and the effects of residual stresses. For these initial investigations, we utilized Off-The-Shelf parts to increase statistical reliability, improve our ability to train our Machine Learning Model and refine our experimental protocol while keeping costs within the budget for the research. In the second phase of this research, we will use MAM parts.

Acoustic shearography technique is a wave-based shearography method. In acoustic shearography, acoustic or ultrasonic waves are used to provide stress loading in the testing materials. Due to the deep penetration of acoustic waves and the flexibility of wave focusing control, acoustic shearography technique can detect deep subsurface defects in solid materials at the optical imaging speed. In this paper, the working principles, and the system setup of the acoustic shearography system are to be presented. Various acoustic excitation methods including acoustic excitation with dispersive acoustic transducers, directional and focusing acoustic wave transducers and pulsed laser beams are discussed. The applications of acoustic shearography for imaging of defects in thick metal structures are demonstrated. The acoustic shearography technique provides a feasible and practical solution for fast NDT of thick and large area structures.

In-situ thermal stress determination in structures is a challenging experimental mechanics task, especially if it requires a nondestructive approach. Thermal stress measurement and management of continuous welded rail (CWR) have become more important for railroad maintenance. 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). This paper examines the potential of the local resonances to provide an estimation of axial stress in rail structures. The local resonances are generated by bonding a piezoelectric element on the rail samples. A 610-mm rail sample was tested, and different axial stress levels were applied by an MTS tensile-compression machine and by measuring the local resonance signature in selected frequency bands to study sensitivity to axial stress of local resonances. The results show that appreciable sensitivities of the local resonances are found under varying stress levels and can be further utilized for in-situ thermal stress determination for rails.

Rail is one of the key elements of the railway system, and its role is to transmit the wheel load to the track bed and guide the train cars along the track. Rail is susceptible to rolling contact fatigue and wear due to being repeatedly subjected to the moving load of the train. This can eventually result in broken-rail damage and train derailment, which if happens on a railroad bridge, it can severely damage the bridge, such as the structural failure of the Tempe Town Lake steel railroad bridge in July 2020 that costed $11 million to repair. Therefore, early detection of defects in rail-bridge system may prevent a critical accident with irreversible damage. The objective of this paper is to use classification-based machine learning techniques to detect broken-rail damage in an open-deck railroad bridge by measuring its acceleration response under the moving load of the train for different speeds. For this purpose, the two-dimensional Finite Element (2D FE) model of a given railroad bridge is created using OpenSEESPy package, which is a Python-3 interpreter of OpenSEES. The changes in the acceleration response due to the damaged rail compared to the undamaged (healthy) rail are characterized by using the Hilbert-Huang Transform in both the time and frequency domains and quantified by defining energy and phase damage indices. The data collected from the 2D FE model are used to train and test several machine learning (ML) classifiers including the Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Decision Tree (DT) algorithms. The results from the data-analytic study show an acceptable level of precision of these classifiers in identifying the damage to the rail-bridge system.

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.

Structural Health Monitoring (SHM) is an approach in which damage detection techniques are utilized to evaluate critical civil infrastructures including bridges, wind turbines, buildings, and tunnels. Typically, non-destructive methods and sensors embedded in and attached to the structures are used to collect data for data interpretation and condition assessment by experts. SHM is very important as it is a critical tool in evaluating the safety and performance of existing structures to prevent accidental failures of structures. The objective of this paper is to investigate the effect of different damage types on the stiffness and fundamental frequency of a bridge model, using a laboratory train model to develop experimental data. This experimentation assumes a constant train speed (0.225 m/s) and a constant train mass (2.07 kg) in all experiments. Considered variables are i) damage type and ii) damage location. Three damage types are i) complete damage (simulated by removal of a bridge member in the model), ii) partial damage (simulated by replacing a regular member with a softer member), and iii) minor damage (simulated by loosening the screws at a joint). Observation parameters include i) sensor location (AB locations) and ii) axial force response spectrum (0 Hz ∼ 0.5 Hz) using fast Fourier transform (FFT). From our experimental data, it was found that the removal of a railway bridge member leads to the reduction of the bridge stiffness and its fundamental frequency.

The research objective is to evaluate the texture conditions in dense-graded asphalt concrete. The research scope is to collect various plant-mixed DGAC mixtures, compact mixtures to slabs, polish the slabs, and evaluate the texture. The testing slabs were polished step-wisely by the wheel-tracking polishing device. WTPD runs in a circle of 284mm in diameter and wears off the surface textures through a standard of procedures, namely 0, 3,000, 6,000, 12,000, 25,000, 50,000、100,000 rounds of polishing efforts. The circular track meter or CTM in lieu of ASTM E2157 was employed to assess the surface texture, non-destructively. The measurement of macro- and micro-texture of the slab surface is the mean profile depth or MPD and as such is a highly accurate yet repeatable texture data, due to the one-micron resolution laser source equipped with CTM. In this study, four DGAC mixtures collected from different regions were examined. MPD data decrease in the first thousands of polishing rounds, due to possible local compaction induced by the repeated wheel loading. MPD data thereafter increase to a maximum between 12,000 to 50,000 polishing rounds and start to decrease. It is possible that the asphalt mixtures were worn off and the texture of aggregates activate to involve in the overall texture condition. More polishing rounds and more repetition tests are recommended to gather more information to be analyzed in the future.

Corrosion of steel reinforcement in concrete highway bridges due to carbonation and chloride attack is an ongoing and prevalent problem. If left undetected in time, late-stage corrosion can lead to section loss of steel rebars, an uneven distribution of internal stresses, and ultimately lead to surface cracks and spalling of the concrete. Nondestructive testing and evaluation (NDT/E) sensors like ground penetrating radar (GPR) are commonly used to detect subsurface objects and anomalies (e.g., concrete cracking, and rebars). In this paper, the feasibility of a 1.6GHz GPR device for detection of steel rebar corrosion is tested in both laboratory reinforced concrete (RC) specimens and in-situ RC structures. For this purpose, three RC specimens (12 x 12 x 5 in³ ) were cast with a No.5 steel rebar (5/8” diameter) at the center of each specimen. Two of the RC panels were corroded using the accelerated corrosion test (ACT) to achieve different levels of corrosion, while the third one was left un-corroded to serve as an intact baseline in our measurements. The RC specimens were kept in a temperature-controlled environment (73 ∼ 77◦F) for six years (2017-2023). In addition, a bridge column of a RC highway bridge underpass (Chelmsford, MA), which exhibits the signs of steel rebar corrosion, was selected for collecting in-situ GPR scans. A known intact RC bridge column was chosen and scanned to serve as the baseline for comparison. In both scenarios, B-scan images were developed from the lab RC specimens and the damaged bridge column. The changes in the reflection amplitudes of the hyperbolic-shaped rebar reflections in GPR B-scan images due to corrosion were studied in both the time and the frequency domains, and cross-validated with the results from previous research. From our experiments, it was found that a 1.6GHz GPR sensor can successfully detect and distinguish corrosion level in three RC specimens and one bridge column by the combined use of 1D and 2D analytic methods.

Evaluating the underside is an important component of inspection, maintenance, and overall sustainability efforts for bridges. Underside inspection and evaluation of bridges is technically challenging due to difficulties with access, lack of customized nondestructive evaluation and health monitoring tools, and lack of well-established understanding of failure modes and degradation mechanisms. A critical problem is the spalling and delamination of concrete bridge decks, which can lead to dangerous conditions where concrete chunks may fall onto under passing vehicles. Conventional human inspection methods can be difficult, expensive and dangerous because gaining access to bridge deck undersides often requires the use of specialized equipment, such as snooper trucks, placing the inspectors at elevated and awkward poses, and redirecting traffic. UAS systems offer an attractive supplement to conventional methods. A UAS can access the underside of bridges and collect inspection and evaluation information, while the inspectors remain on the ground. This paper describes research towards developing UAS systems that can navigate on the underside of bridges and then collect image data, evaluate the image data with automated artificial intelligence methods, and use onboard nondestructive evaluation techniques including noncontacting microwave radar, and contacting acoustic response methods. Results from experiments, data analysis and signal processing will be presented.

A nondestructive testing method using piezoelectric cement sensors (PEC sensors) to monitor the free chloride ion content of hardened concrete by using the electromechanical impedance (EMI) technique is proposed. Piezoelectric cement, a sensing element, is a 0–3 type cement-based piezoelectric composite comprising 50% lead zirconate titanate (PZT) particles. The rapid chloride ion permeability test (RCPT) complies with ASTM C1202 to accelerate the content and penetration of chloride ions in concrete while measuring the electromechanical impedance spectrum of concrete with a PEC sensor. The free chloride ion content in concrete was determined following ASTM C1218. The conductance root-mean-square deviation (GR) was calculated within the effective frequency range to assess chloride ion (Cl) in concrete. The results indicate that the total passed charge, according to the RCPT specification, lies in the linear to nonlinear transition region of the passed charge and free chloride ion content curve. The conductance within the effective frequency monitored with the PEC sensor increases with the free chloride ion content in concrete. The effective frequency band is 1350–2000 kHz. The Cl-GR curve of concrete having an exponential relation is established. After chloride ion penetration in RCPT for about 6 hours, the GR value will be increased sharply. The PEC sensor has the ability to monitor free chloride ion content in concrete, and the conductance RMSD correlates highly with the free chloride ion content. The proposed approach can be used to monitor the free chloride ion content of concrete structures.

Presence of chloride ions is the main cause responsible for the corrosion of steel rebars and other components in concrete structures. Detecting the presence and amount of chloride ions in concrete, however, is a challenging task. In this paper, we present our development on the use of synthetic aperture radar (SAR) imaging for the remote detection of chloride ion content in concrete. SAR imaging is a remote sensing technique capable of performing noncontact subsurface inspection of dielectric materials like Portland cement concrete. Such techniques can be applied to field inspection of concrete structures and laboratory material characterization. The objective of this paper is to demonstrate the application of SAR imaging on characterizing chloride content inside oven-dried concrete specimens, using a 10-GHz central frequency SAR imaging sensor. Twelve concrete specimens (0.3x0.3x0.05 m3 ) with a 0.45 water-to-cement ratio were manufactured in six groups of different chloride contents (0%, 2%, 4%, 6%, 8%, and 10% of cement weight). Four SAR image-based parameters were developed from each SAR image of concrete specimens for chloride, including integrated amplitude Iint, average maximum amplitude Imax, critical contour area Ac, and average Gaussian curvature of critical contour Kavg. A parametric analysis on the combined use of different numbers of SAR image parameters was carried out to determine the optimal application of SAR image parameters on the chloride detection problem in this paper. From our result, it is found that, for the purpose of chloride detection inside concrete specimens using single SAR parameter, integrated amplitude Iint. When using two SAR image parameters, combination of integrated amplitude Iint and critical contour area Ac shows the best performance among other two-parameter combinations. When using three SAR image parameters, combination of integrated amplitude, critical contour area, and average maximum amplitude shows the best performance among other three-parameter combinations. Furthermore, it is interesting to report that combination of three SAR image parametersprovides the overall best performance among all other combinations for the chloride detection problem in oven-dried concrete specimens.

Early-age concrete undergoes displacements and volume changes due to ongoing processes such as settlement, hydration, shrinkage, and cracking, which can strongly affect its durability and long-term performance. In this paper, fresh concrete is monitored by the non-destructive techniques of Acoustic Emission (AE) and Digital Image Correlation (DIC). Elastic waves released by the physical processes taking place while concrete is in a fresh state can be well-recorded by AE, while the three-dimensional strain and displacement evolution on the surface can be measured by DIC. Monitoring fresh concrete is of paramount importance to ensure the desired final mechanical properties, especially when novel admixtures for internal curing such as SuperAbsorbent Polymers (SAPs) are added to the mixture. SAPs are particles that can swell by absorbing water when exposed to it, and later release it back to the cementitious matrix when the internal relative humidity linked to the capillary pressure decreases, mitigating autogenous shrinkage. These admixtures strongly interact with the microstructure, resulting in an increased amount of AE activity. The motivation of this study is to obtain real-time information on the different ongoing processes in fresh concrete using AE and compare the results to concrete containing SAPs. Specimens are subjected to different environmental conditions, to monitor the changes in the SAP activity. Results are complemented by DIC to confirm the mitigation of shrinkage by the SAPs. The DIC results showed that SAPs mitigate settlement and shrinkage in early-age concrete, while AE showed SAP concrete exposed to windy conditions demonstrated a delay in the SAP activation, lower amplitude values and higher peak frequency values than the ambient SAP concrete.

The adoption of self-healing cementitious materials has gained attention as an alternative to costly and labour-intensive manual repairs. Cementitious blends possess an inherent ability to repair formed cracks through so-called autogenous healing. Whereas the efficiency of autogenous healing remains limited as moisture needs to access the cracks, the healing capacity can be improved through the inclusion of superabsorbent polymers (SAPs). To encourage the use of these self-healing blends within the construction industry, an assessment of the healed state is necessary to ensure a structure’s safety. The requirements for such evaluation method comprise the ability of assessing the regained mechanical performance, while maintaining the structural capacity of the member under study. A non-destructive method that has proven its potential is the application of ultrasonic waves, which are sensitive to the elastic properties of the material they travel through. Coupled ultrasound is currently most often used, while air-coupled ultrasonic measurements allow to reduce the occurring coupling variability. In this study, the self-healing evolution of cementitious mixtures with and without SAPs was assessed through coupled and air-coupled ultrasound. A comparison between both techniques confirmed the potential of air-coupled ultrasound, paving the way for automated self-healing evaluations.

Cracking of cementitious materials affects the durability of concrete structures and might lead to premature failure. As manual repairs are costly and labor-intensive, self-healing mixtures have been studied. The advantage of cementitious blends lies in the inherent ability of the material to repair damage through autogenous healing. As water is essential to be present to induce autogenous healing, the healing ability can be improved by adding water reservoirs in the form of superabsorbent polymers (SAPs). As a wide variety of SAPs with different characteristics exists, an assessment of their capacity to improve the self-healing ability is necessary to optimize the mix design. While most standardized evaluation techniques are limited in their characterization potential or due to their intrusive nature, ultrasonic measurements allow for a non-destructive material characterization. Due to their sensitivity to the obtained microstructure, the damage present and the elastic properties of the material under study, the self-healing evolution can be monitored, and the results provide information on the regained mechanical performance. In the present study, various set-ups are utilized to assess the self-healing capacity of mortars with and without SAPs. The experimental framework includes coupled ultrasonic evaluations through surface wave and transmission measurements. In addition, numerical simulations were performed to isolate the healing layer and simulate the effect of healing by increasing the stiffness of the material in the crack. A comparison between experiments and simulations allowed to assess the elastic modulus of the deposited healing products.

Currently, ultrasonic testing (UT) is widely used in concrete for identifying the sub-surface or surface flaws. However, most the UT can only provide the qualitive evaluation of the flaws. Tomography technology is capable to visualize the damage and provide its positioning information in concrete by reconstructing the ultrasound. For example, the slurry leakage state in the grouting sleeve, and the location of the inclusion in the concrete pile foundation can be displayed by ultrasonic tomography. However, concrete material is highly heterogenous due to its complex components (aggregate, mortar, internal void). All those complexities can cause significant impact on the tomographic results. Especially for the aggregate, sometimes its dimension is very closed to the testing objects. It greatly affects the recognition and location of defects by ultrasonic testing. Therefore, this research proposed to reveal the influence of aggregate, defect size and the effect of type, tomographic pixels on tomographic images in concrete. An optimum transducer arrangement and tomography algorithms in terms of ray-trace method was proposed to achieve the high accurate and resolution of tomography. Finally, the comparison of tomographic images between the imaged location and the embedded location is evaluated and then tomographic states is assessed accordingly.

Engineering structures made of different materials have an ancient root in many engineering disciplines such as civil, mechanical, and aerospace engineering. Such structural systems or structures are commonly known as composite structures or composite systems. In civil engineering, reinforced concrete (RC), prestressed concrete (PC), and concrete-FRP (fiber reinforced plastic) systems are three well-known examples. Different from RC and PC structures, concrete-FRP systems are usually formed by using FRP sheets/plates/strips to externally bond to the surface of damaged concrete structures for retrofitting, repair, and strengthening. Externally bonded FRP sheets/plates/strips can provide additional tensile strength, or shear strength, or compressive strength (through the confinement effect) to existing concrete structures. However, air pockets and FRP delamination introduced during improper installation can lead to the brittle, immature failures of concrete-FRP systems. Furthermore, the surface information of concrete structures for condition assessment becomes unavailable after the installation of FRP. In this paper, a review of nondestructive evaluation (NDE) techniques for concrete-FRP systems is provided. Contact and non-contact NDE techniques are compared on their detection of subsurface anomaly inclusion/ delamination in concrete-FRP systems. An emphasis is placed on the recent development of remote NDE techniques such as laser Doppler vibrometer (LDV) and synthetic aperture radar (SAR). Challenges in the condition assessment of concrete-FRP systems in the field are also discussed for different NDE techniques.

Aging infrastructure worldwide has led to an interest in using innovative solutions for non-destructive damage assessment. The traditional in-person assessment of large-scale infrastructure like dams and bridges etc. can be very time and resource consuming. With advances in technology, image processing and machine learning have shown promise in providing alternate ways of damage assessment in large-scale and difficult-to-access infrastructure. However, this approach has mostly been applied passively and not in real-time. The work presented here describes a supervised machine learning based approach to damage analysis of concrete structures. Python programming language is used to write and train algorithms to provide a real-time damage analysis. In addition to crack detection, the dimensional analysis provides additional information regarding an existing or developing crack. With this type of real-time information, timely action can be taken regarding decisions like performing repairs or decommissioning a structure for public safety.

In this work, thermoplastic composite materials, potentially usable for applications in the transport field, have been prepared starting from a commercial PLA/PBAT blend including calcium carbonate as a filler and a commercial jute fiber fabric. Specimens of adequate size cut from hot pressed plates and subjected to flexural tests were systematically analyzer with the aid of an optical non-destructive technique to highlight the dissipative mechanisms induced by the test and responsible for the failure of the material. The NDE technique adopted belongs to the optical methods in full-field and non-contact modality.

Electromechanical properties of carbon fibers enable non-destructive evaluation (NDE) of carbon-fiber-reinforced plastic (CFRP) structures by monitoring electrical resistance in real-time. This NDE technique is named as ‘self-sensing’, as it employs the material’s intrinsic features like human nerves. This technique was applied to evaluate the size of damage in CFRP samples. Electrical resistance measured in real-time during machining increasing size of concentric circles was analyzed, and polynomic correlations were identified. To ensure reliability and take uncertainties account, probability-based tools, Markov chain Monte Carlo and Bayesian algorithm, were applied. The potential applicability of the established system to repeated impact loads, considering damage progression due to unexpected strikes in real applications, was also verified.

Rapid-aging civil infrastructures have become social issues worldwide. Developed countries have faced difficulties e.g. cost increase in infrastructure maintenance, shortage of highly-skilled human resources. Thereafter, infrastructures have been inspected infrequently e.g. every 1~10 year. 3D mapping powered by light detection and ranging (LiDAR) platforms has recently been brought to public attention. In infrastructure maintenance, LiDAR having high-precision and long-measurement range is expected to enhance inspection frequencies and lower the maintenance cost. Time-of-flight (ToF) approach is generally utilized for long-range measurement, however it suffers from low depth resolution. Although development of dual-comb interferometry enabled high-precision ToF rangefinders, sophisticated stabilization is required for two lasers. On the other hand, frequency-modulated continuous-wave (FMCW) LiDAR can perform ranging with a single laser. However, its measurement range is limited by its receiver bandwidth. Although FMCW LiDAR with low-speed analog-to-digital converters can conduct long-distance ranging by reducing its wavelength-sweeping rate, the measurement refresh rate is sacrificed. In this article, we propose a long-range FMCW LiDAR employing wavelength-swept optical frequency comb which overcomes ranging limitation caused by its receiver bandwidth. In addition, the spectrum of such optical frequency comb is apodised to circumvent its inherent ranging ambiguity. We have performed proof-of-concept experiments using the proposed LiDAR. With optical fibers for emulating beam propagation, we have succeeded in ranging corresponding to 1,605-m free-space propagation with a 10-MHz FMCW receiver. Our proposal will make an important contribution to infrastructure maintenance.

Optical fiber strain distribution sensors have become an excellent tool for Structural Health Monitoring (SHM). However, the fiber’s fragility and the time-consuming installation process have slowed down the technology’s broad application. In this study, we embedded optical fibers with textiles, allowing a much faster installation process and the ability to design specific sensor patterns for multiaxial sensing. Specifically, the fiber attached to the textile was embedded, forming a rosette configuration. The textile was bonded to an aluminum cantilever and subjected to different loads. The strain distribution from the optic fiber was measured using an Optical Frequency Domain Reflectometry (OFDR) instrument. The fiber sensor response was compared with Strain Gauges to determine the reliability of the fiber sensor.

Sandwich composites are vulnerable to low-velocity impacts that they are mainly exposed to. But non-destructive testing for sandwich composite is troublesome due to its discrete and inhomogeneous nature. This study investigates the damage severity of the non-conductive foam core sandwich composite using the electrical resistance change of Carbon fiber-reinforced plastics skins under impact. Four severity levels are paired to form 4C2 binary classifications, Single Linear Discriminant Analysis (LDA), using electrical resistance change as an input. By combining six Single LDAs’ probabilistic results into one system, this system can represent the specimen’s current state intuitively and diagnose the damage severity synthetically.

Micro-resolution ultrasound is a process that can characterize materials using a focused beam immersion ultrasonic probe, an infrared (IR) laser vibrometer, an immersion tank, and a sample holder. The focused beam immersion ultrasonic probe has a spot size of 0.3mm with 20MHz at a focal point of 28mm. The focal diameter of the IR laser can be as small as 10µ, allowing for a higher probability-of-detection (POD) for defects in materials. A 3cm-by-3cm sample of molded fiber glass, similar to what is used in civil aviation aircraft was examined using both IR and HeNe lasers. A system interface using Python and MATLAB was used to automate a nondestructive evaluation process that produced microscopic resolution of the control sample (a grade-91 steel plate) and two samples of composite pieces from a Cirrus Aircraft. In conclusion, this study showed that microscopic resolution can be achieved with less sample preparation using an IR laser, in place of a HeNe laser that is traditionally used.

There is a growing need for non-invasive structural health monitoring in extreme environments. For nuclear power plants, pressure and temperature sensing under hazardous environment plays an important role for coolant system safety and stability management. Current sensing methods are intrusive, and suffer from degradation in the plant environment, limited life cycle, and complicated repair and replacement procedures. In this paper, we present an advanced Bi-In-Sn liquid metal (LM) transducer with the addition of candle-soot nanoparticles (CSNP) for improved photoacoustic efficiency and a metallic stencil for control of the liquid metal layer thickness. The sensitivity of the liquid metal candle-soot nanoparticle (LM-CSNP) ultrasound transmitter was characterized under 2 mJ/cm² at 65 °C, and 6 mJ/cm2 at 100 °C —300 °C. Compared with existing LM transmitter, the newly presented transmitter showed a sensitivity 6.6 times stronger than previously reported LM only transmitter.

This paper is based on the continuous wavelet transform to achieve modal parameter identification of the response signal of a wing under environmental excitation. Suitable wavelet basis functions are selected for the characteristics of flutter data to achieve a good time-frequency analysis of flutter data. And the endpoint effect is effectively suppressed by the predictive extension method of support vector machine. The identification of wavelet ridges and the wavelet cross-section are obtained through a crazy creep algorithm to achieve the identification of modal damping in the chattering data. Finally, the method is applied to the wind tunnel test data of a 3d-printed wing to identify its modal parameters, and is compared with its numerical simulation data to verify the reliability of CWT applied to the identification of modal parameters of wing flutter data under environmental excitation.

In this paper, we propose a technique for classifying different types of damages in honeycomb composite sandwich structures (HCSS) using guided wave-based structural health monitoring (GW-SHM) systems that can work well even when data loss is lost for chunks of time. Although damage classification is important for deciding the course of action for usage, repair, and replacement, we show that there is overlap in the amplitude characteristics between different types of damages. This problem can be further exasperated when data is lost in chunks due to attenuation, physical damage, electromagnetic interference, and hardware faults. Reliable signal models are not always available for interpolation-based data recovery in such cases. First, we simulated the loss of 10% to 50% samples in experimentally collected data and recovered the signals using orthogonal matching pursuit with an error consistently below 0.1%. Next, we extracted ten features from the recovered signals in both time and frequency domains. We eliminated four features based on correlation analysis to improve the classification performance. We tried multiple classifiers to distinguish between healthy structure and four types of damages: lost film adhesive, teflon release film, core crush, and high density core in HCSS, and obtained perfect classification accuracy with random forest classifier and an optimal feature set based on feature importance. Due to smaller size and computational efficiency, these models are best suited for edge implementation and in situ monitoring. Also, models using these statistical features are portable to other structures as they are independent of material properties.

In order to predict the flutter boundary of the wing under the action of atmospheric turbulence, the obtained turbulence signal is first denoised, and the attenuation signal containing a single mode is obtained by using the variational modal decomposition method. The data containing few attenuation points are extended by using machine learning methods, and the matrix pencil method is used to identify the modes. Finally, the stability criterion is calculated by using the modal parameters, The flutter boundary is calculated from these data. In this paper, the wind tunnel test data are used to analyze and solve the modal, and the flutter boundary is accurately predicted in advance.

A flutter boundary prediction method based on HHT, and machine learning is proposed to predict the flutter velocity before the wind speed reaches the subcritical state. Natural excitation technique is used to extract impulse response signals. EMD (empirical Mode decomposition method) is used to decompose the signal. Hilbert spectrum was obtained and analyzed by HHT to decompose the signal. The analysis methods included HHT spectrum and marginal spectrum analysis, so as to extract the characteristic quantity and establish the classification model according to different flight states. Then, regression models were established under different flutter modes for flutter degree analysis. During the prediction, according to the classification performance of the data to be measured, the flutter degree analysis result is weighted to obtain the flutter degree corresponding to the current wind speed, and then the flutter wind speed is calculated. In the selection of machine learning algorithm, naive Bayes algorithm, K-nearest neighbor algorithm and other machine learning algorithms are used to construct the classification model, linear regression, Gaussian process regression and so on are used to construct the regression model. The results show that the K-nearest neighbor algorithm performs best in the classification algorithm, while the Gaussian process regression algorithm performs best in the regression algorithm. Through the cross-validation of the test data, the proposed method can accurately predict the critical flutter velocity when it is far away from the flutter boundary through flutter mode recognition and flutter degree analysis.