Simulated Delamination Research Articles

Ensemble learning model for concrete delamination depth detection using impact echo

Delamination in concrete structures typically arises from the expansion of steel bar corrosion and the separation of the bonding layer. The presence of these delamination defects leads to a reduction in overall structural integrity, safety, and longevity. However, the delamination depth localization remains challenging. To address this issue, this paper introduces a predictive approach that combines time-frequency features extracted from vibration signals with ensemble learning models. The innovation presented in this paper can be summarized in two key aspects: (1) Signal time-frequency spectrograms are transformed into image features using the Histogram of Oriented Gradients (HOG) method, which serves as input attributes. The utilization of ensemble learning models facilitates both rapid and precise depth estimation of delamination signals, especially those with shallow defects. (2) By analyzing the time-frequency spectrogram of the vibration signals, a double exponential model is established between the depth-to-thickness ratio of delamination and the pixel matrix Eulerian distance. This may provide a new perspective for the depth estimation of delamination defects whose corresponding main frequencies are difficult to locate due to the influence of geometry and boundary effects. Laboratory experiments were conducted on concrete slab with simulated delamination defects at different depths to confirm the proposed method's effectiveness.

Deep learning-based prediction of delamination growth in composite structures: bayesian optimization and hyperparameter refinement

Abstract In this study, a novel method for detecting the growth of delamination in sandwich structures has been proposed. To this end, we suggested hybridizing the Deep Learning techniques (DL) and Finite Element Method (FEM) for predicting the growth of delamination in this structures. A dataset of simulated delamination growth under different delamination sizes has been produced using the FEM method. Then, a DL model has been trained using this dataset to precisely predict the growth of delamination. This study focused on predicting delamination growth using a tuned and optimized deep learning based regressor. Therefore, to find the ideal set of hyperparameters, the Bayesian optimization algorithm has been used for selecting the best structure and enhancing the regressor performance. Afterward, the model was evaluated and multiple processes were conducted to improve its behavior and solve its stability and overfitting issues. Particularly, an inconsistency between validation loss and training loss has been initially detected in the behaviour of the model, which may indicate overfitting. To tackle this issue, dropout regularization has been added, which improved the consistency between the loss functions but results in less smooth convergence from the expectations. So, in a third study, dropout and L1 regularization has been combined to improve the stability of the model. This combination achieved a consistent and smooth convergence between the validation and training loss functions. The findings highlight the importance of hyperparameter optimization and regularization techniques in improving regression model performance. The study shows the efficiency of Bayesian optimization in hyperparameter tuning and the iterative optimization of a regression model. Furthermore, the outcomes show that the suggested method can identify and predict delamination growth with high accuracy.

Damage localization using contact and non-contact narrow frequency band elastic wave generation

In this paper analyses of non-contact elastic wave generation based on an air-coupled transducer (ACT) are presented. Authors compared contact piezoelectric based (PZT) wave generation, with non-contact one using ACT for a narrow frequency band excitation. Elastic wave sensing was based in both cases on a scanning laser Doppler vibrometer. The problem of ACT slope angle for the effectiveness of elastic wave modes generation is analysed. It was shown that for an excitation frequency of 40 kHz it is possible to generate A0 wave mode both in the case of PZT and ACT. Both methods of wave generation are compared based on an analysis of damage localisation results. For this purpose wave irregularity mapping (WIM) algorithm was utilised together with two attenuation compensation methods (based on energy normalisation and weighting function). Specimens in the form of simple square GFRP panel and bonded GFRP aerospace structure with omega stiffeners were investigated. The authors investigated simulated damage (Teflon inserts) and real delamination. In this research very cheap and popular, low-frequency ACT was utilised giving the possibility of realisation of low-cost excitation method. ACTs connected with a laser vibrometry sensing allows of realisation of full non-contact damage localisation method. Authors presented also non typical but very useful application of SLDV allowing to visualise the effect of acoustic wave generation in the air and determination of the pattern of generated acoustic waves. This was based on the photo-acoustic effect.

Data Processing Scheme for Laser Spot Thermography Applied for Nondestructive Testing of Composite Laminates

This paper proposes a data processing scheme for laser spot thermography (LST) applied for nondestructive testing (NDT) of composite laminates. The LST involves recording multiple thermographic sequences, resulting in large amounts of data that have to be processed cumulatively to evaluate the diagnostic information. This paper demonstrates a new data processing scheme based on parameterization and machine learning. The approach allows to overcome some of the major difficulties in LST signal processing and deliver valuable diagnostic information. The effectiveness of the proposed approach is demonstrated on an experimental dataset acquired for a laminated composite sample with multiple simulated delaminations. The paper discusses the theoretical aspects of the proposed signal processing and inference algorithms as well as the experimental arrangements necessary to collect the input data.

Damage imaging method for composites laminates based on sparse reconstruction of single-mode Lamb wave

To address the problems of anisotropy and mode mixing in damage imaging on anisotropic composites laminates using ultrasonic Lamb wave, a sparse reconstruction imaging method based on S0 single-mode is proposed in this paper. In this approach, the sparsely arranged piezoelectric ceramic drivers sensor group on the surface of the measured composite laminate is employed to excite and receive Lamb wave. The received signal shows sparse under a pre-built damage dictionary with damage scattering signal waveform, and the damages can be localized and imaged using a sparse reconstruction method. Considering the velocity difference between S0 and A0 mode of the received Lamb wave signal, the S0 mode is extracted by intercepting the signal before the peak of the first wave packet to eliminate the influence of Lamb wave mode mixing. In order to minimize the influence of anisotropy of composites laminates on Lamb wave propagation, the Lamb wave propagation direction is divided evenly into several sectors, and Lamb wave propagation direction within a certain sector is considered as propagation in a quasi-isotropic material. The sparse reconstruction is solved by using the basis pursuit de-noising algorithm, which can find the optimal solution with a small number of iterations in noise interference. The experiments on composite fiber laminate specimens show that the proposed method can accurately locate simulated single delamination damage and has less artifact interference compared to the delay-and-sum method.

Edge delamination width prediction of 3D body‐in‐white part by finite element‐based corrosion simulation and neural networks

AbstractThe edge corrosion simulation of Kapfer et al. is used to extract time‐dependent delamination widths of metal sheet edges with different spatial orientations, encoded by edge classes and angular displacements with respect to the ground. The corrosion behavior of the edges of a body‐in‐white part is examined with respect to the spatial orientation and compared to the simulated delamination widths, demonstrating that the edge corrosion simulation is also able to model more complex structures. Since scaling the finite element method‐based corrosion simulation to the whole part requires a lot of memory and computational time, different neural network types are evaluated to predict the delamination widths. Together with the geometric properties of the part edges, a fast and accurate prediction of edge delamination of body‐in‐white parts is possible, requiring few time‐consuming corrosion simulations only in the training phase of the networks.

Ultrasonic Structural Health Monitoring Approach to Predict Delamination in a Laminated Beam Using d15 Piezoelectric Sensors

Abstract Ultrasonic structural health monitoring (SHM), employing embedded piezoelectric elements to actuate and sense ultrasonic waves, has greatly advanced in recent years. This paper presents a novel approach to address the prevailing challenges in the inspection of laminated structures for delamination using shear-mode (d15) piezoelectric transducers, composed of lead zirconate titanate (PZT). To experimentally evaluate the effectiveness of the proposed approach, a beam-like laminated specimen consisting of internally embedded d15 square PZTs was fabricated with simulated delamination at the interface of an adhesive joint. The evaluation of the results showed that the location of shear-mode actuators is a critical factor to detect delamination and to predict the propagation path of delamination. Delamination initiated close to actuators is more likely to be detected owing to their remarkable sensitivity of structural stiffness surrounding their region. The antisymmetric A0 wave mode generated by these actuators exhibits high interaction with damage, suggesting internally embedded d15 PZTs are a viable approach that can potentially advance the inspection tools of ultrasonic SHM.

Numerical and experimental investigation of delaminations in a unidirectional composite plate using NDT and SHM techniques

In this paper, the non-destructive testing (NDT), structural health monitoring (SHM), and scanning laser Doppler vibrometer (SLDV) techniques were presented to quantify three simulated delaminations inserted at different depths of a unidirectional composite plate. First, the RollerFORM and Omniscan equipment were sufficiently used to identify the delaminations. Second, in conjunction with guided waves, the developed imaging method was successfully used to detect and quantify the interested delaminations. The tuning curves were determined experimentally to define the dominant Lamb wave modes of incident waves. Third, multi-physics three-dimensional finite element simulations of propagating and interacting Lamb waves with delaminations were implemented to extract the wavefield data for wavenumber analysis. The experimental part was conducted to validate the numerical results using SLDV. The effect of the delamination depth on the trapped waves generated over the delamination region was studied numerically and experimentally. The results showed that trapped waves could be affected by the delamination depth. Both numerical and experimental results demonstrated that the near surface delamination has strong trapped waves over the delamination region while the far surface delamination has weak trapped waves. The energy distribution maps of numerical and experimental wavefields data sufficiently quantified the interested delaminations. A good agreement was achieved between the numerical and experimental results.

Delamination process in cross-ply UHMWPE laminates under ballistic penetration

Cross-ply unidirectional laminates made of ultra-high molecular weight polyethylene fibers are widely used as components of bullet-proof vests. To simulate the delamination process of the material under penetration, we constructed a numerical mechanical model, which was validated by tests using 7.62 × 39 mm rifle bullets penetrating laminates of different thicknesses. The results show that the delamination region is essentially diamond-shaped. The simulated delamination region is in good agreement with the experimental data. It is found that the delamination region increases with the compression modulus along the fiber direction. The delamination region increases when the shear strength between the fabric layers decreases; However, it is little influenced by the normal strength. The delamination region of the front face of the laminate is little influenced by the failure strain of the material and the initial velocity of the bullet. The delamination region of the back face increases with the failure strain and decreases with the initial velocity.

Ultrasonic detection and characterization of delamination and rich resin in thick composites with waviness

A multi-frequency ultrasonic method was proposed to detect and characterize delamination and rich resin in thick composites with waviness. Addressing the challenges for ultrasonic inspection caused by waviness, multi-layer structure, and multiple defect types, ultrasound propagation in thick wavy composite was investigated through a dedicated numerical model built in OnScale®. Key features of ultrasonic testing and composite sample, including water immersion environment, fiber waviness, uneven inter-ply resin distribution, and side-drilled hole (SDH)-simulated delamination, were embraced in the model. Via this model, waviness and thick resin layers were found to cause disturbance and wave vector deviation to inter-ply reflection signal and thus introduce difficulties for delamination detection based just on the signal to noise ratio. In addition, reflections from inter-ply resin and SDH were revealed with different reliance on probe frequencies. Using a 5 MHz probe with time corrected gain, ultrasonic testing in wavy composite was performed experimentally. Based on the reliance of defect characterization on frequency, SDHs and rich resin were differentiated experimentally in the B-scan images with various filtering frequencies, demonstrating the effectiveness of the proposed method for detection and characterization of delamination and rich resin in thick wavy composite structures.

Multimode Guided Wave Detection for Various Composite Damage Types

This paper presents a new methodology for detecting various types of composite damage, such as delamination and impact damage, through the application of multimode guided waves. The basic idea is that various wave modes have different interactions with various types of composite damage. Using this method, selective excitations of pure-mode guided waves were achieved using adjustable angle beam transducers (ABTs). The tuning angles of various wave modes were calculated using Snell’s law applied to the theoretical dispersion curves of composite plates. Pitch–catch experiments were conducted on a 2-mm quasi-isotropic carbon fiber-reinforced polymer (CFRP) composite plate to validate the excitations of pure fundamental symmetric mode (S0) and shear horizontal mode (SH0). The generated pure S0 mode and SH0 mode were used to detect and separate the simulated delamination and actual impact damage. It was observed that S0 mode was only sensitive to the impact damage, while SH0 mode was sensitive to both simulated delamination and impact damage. The use of pure S0 and SH0 modes allowed for damage separation. In addition, the proposed method was applied to a 3-mm-thick quasi-isotropic CFRP composite plate using multimode guided wave detection to distinguish between delamination and impact damage. The experimental results demonstrated that the proposed method has a good capability to detect and separate various damage types in composite structures.

Evaluation of Delamination in Concrete by IE Testing Using Multi-Channel Elastic Wave Data.

The main objectives of this study are to develop a non-destructive test method for evaluating delamination defects in concrete by the Impact-echo test using multi-channel elastic wave data and to verify the validity of the proposed method by experimental studies in the laboratory. First, prototype equipment using an eight-channel linear sensor array was developed to perform elastic wave measurements on the surface of the concrete. In this study, three concrete slab specimens (1500 mm (width) by 1500 mm (length) by 300 mm (thickness)), with simulated delamination defects of various lateral dimensions and depth, were designed and constructed in the laboratory. Multi-channel elastic wave signals measured on the three concrete specimens were converted to the frequency-phase velocity image by using the phase-shift method. A data processing method was proposed to extract the dominant propagating waves and non-propagating waves from the dispersion images. The dominant wave modes were used to evaluate delamination defects in concrete. It was demonstrated that the surface wave velocity values were useful for characterizing the shallow delamination defects in concrete. In addition, the peak frequency of non-propagating wave modes extracted from the dispersion images gives information on the lateral dimensions and depths of the delamination defects. This study also discussed the feasibility of combined use of the results from propagating and non-propagating wave modes to better understand the information on delamination defects in concrete. As will be discussed, the multi-channel elastic wave measurements enable more accurate, consistent, and rapid measurements and data processing for evaluation of delamination defects in concrete than the single-channel sensing method.

Air-coupled ultrasonic testing to estimate internal defects in composite panels used for boats and luxury yachts

In constructions of boats and luxury yachts, Glass Fiber Reinforced Polymer (GFRP) composite materials are widely employed owing to their high mechanical performance, light weight and flexibility in manufacturing complex shapes. However, GFRP composite materials can be affected by several intrinsic defects (e.g. inclusion, delamination, resin excess or lack), because of both wrong production process and post-production handling. Although conventional ultrasonic testing is the most applied method to enhance the structural safety and quality, it presents a strong limitation in achieving an efficient contact between the transducers and the testing surfaces. Air-coupled ultrasonic testing can be used as a valid alternative, overcoming the previous disadvantage. Aim of this study is to demonstrate the ability of air-coupled ultrasonic testing (ACUT) for qualitatively and quantitatively evaluation of internal defects in GFRP composite panels as those typically used for boats and luxury yachts construction. Simulated delaminations of different sizes have been analysed as well as excess and lack of resin. Results highlighted that a good signal-to-noise ratio can be achieved for delamination greater than 15 mm with correct defect shape recovering, hence making ACUT an appropriate choice for the interactive engineering design and manufacturing of boats.

Modified sparse reconstruction imaging of lamb waves for damage quantitative evaluation

Lamb wave based sparse reconstruction imaging is promising for damage localization in structural health monitoring (SHM) and nondestructive testing (NDT). This method requires accurate dictionary construction to match damage reflected Lamb wave signals. Damage is assumed to be a point-like reflector in previous research, however, most structural damages tend to occupy a physical extent within the structure, which limits dictionary accuracy and thus reduce the performance of sparse reconstruction imaging. In this study, dictionary construction is extended to both point-like damage and damage occupying a physical extent, a modified sparse reconstruction imaging method is subsequently proposed based on the extended dictionary. Compared with conventional method, imaging performance of the modified method is improved due to superior dictionary construction. Another advantage of modification is the ability of forming damage outline for quantitative evaluation. The effectiveness of the modified method is experimentally demonstrated with simulated delamination in composite laminates.

Stress states and crack behavior in plasma sprayed TBCs based on a novel lamellar structure model with real interface morphology

Abstract To ascertain the crack growth behavior and coalescence mechanism in thermal barrier coatings (TBCs) is beneficial for understanding the failure of TBCs and proposing the probable optimization methods. In this work, a novel lamellar structure model with real interface morphology is developed to explore the crack growth behavior and the failure mechanism of TBCs during thermal cycling. Three typical defects which include pore, inter-splat crack, and intra-splat are incorporated in the model. To simulate the oxidation process of the bond coat (BC) realistically, The oxidation growth process is simulated via changing the BC properties to thermally grown oxide (TGO) properties layer by layer. The effects of the lateral growth strain distribution through TGO thickness on the stress states are executed. Moreover, the influences of BC creep on the crack growth and coating lifetime are further elaborated. The results show that the larger the lateral growth strain gradient, the smaller the residual tensile stress. The irregular interface morphology results in the redistribution of residual stresses. Although the pores and cracks can alleviate the tensile stress near the valley, large stress concentration will occur near them. At the early phase of thermal cycling, the cracks grow steadily. After more cycles, the cracks propagate rapidly and merge with others. The simulated delamination path is in agreement with the experiment results. Not only does BC creep change the crack coalescence mechanism, it also decreases the thermal cyclic lifetime of TBCs. The coating optimization method proposed in this study provides another option for developing advanced TBCs with longer lifetime.

Vibration-Based In-Situ Detection and Quantification of Delamination in Composite Plates.

This paper presents a new methodology for detecting and quantifying delamination in composite plates based on the high-frequency local vibration under the excitation of piezoelectric wafer active sensors. Finite-element-method-based numerical simulations and experimental measurements were performed to quantify the size, shape, and depth of the delaminations. Two composite plates with purpose-built delaminations of different sizes and depths were analyzed. In the experiments, ultrasonic C-scan was applied to visualize the simulated delaminations. In this methodology, piezoelectric wafer active sensors were used for the high-frequency excitation with a linear sine wave chirp from 1 to 500 kHz and a scanning laser Doppler vibrometer was used to measure the local vibration response of the composite plates. The local defect resonance frequencies of delaminations were determined from scanning laser Doppler vibrometer measurements and the corresponding operational vibration shapes were measured and utilized to quantify the delaminations. Harmonic analysis of local finite element model at the local defect resonance frequencies demonstrated that the strong vibrations only occurred in the delamination region. It is shown that the effect of delamination depth on the detectability of the delamination was more significant than the size of the delamination. The experimental and finite element modeling results demonstrate a good capability for the assessment of delamination with different sizes and depths in composite structures.

Recent Advances in Piezoelectric Wafer Active Sensors for Structural Health Monitoring Applications.

In this paper, some recent piezoelectric wafer active sensors (PWAS) progress achieved in our laboratory for active materials and smart structures (LAMSS) at the University of South Carolina: http: //www.me.sc.edu/research/lamss/ group is presented. First, the characterization of the PWAS materials shows that no significant change in the microstructure after exposure to high temperature and nuclear radiation, and the PWAS transducer can be used in harsh environments for structural health monitoring (SHM) applications. Next, PWAS active sensing of various damage types in aluminum and composite structures are explored. PWAS transducers can successfully detect the simulated crack and corrosion damage in aluminum plates through the wavefield analysis, and the simulated delamination damage in composite plates through the damage imaging method. Finally, the novel use of PWAS transducers as acoustic emission (AE) sensors for in situ AE detection during fatigue crack growth is presented. The time of arrival of AE signals at multiple PWAS transducers confirms that the AE signals are originating from the crack, and that the amplitude decay due to geometric spreading is observed.

Sparse reconstruction imaging of damage for Lamb wave simultaneous excitation system in composite laminates

Most Lamb wave methods utilize transducer array for damage detection and localization. Simultaneous excitation of multiple transmitters is a strategy for efficient data acquisition, in which matched filtering is generally applied for signal separation from different sources. However, the signal-to-noise ratio (SNR) and subsequent damage imaging performance is limited due to transmitter interference. To obtain both efficient data acquisition and high imaging performance, sparse reconstruction imaging for Lamb wave simultaneous excitation system is proposed in this paper that exploits the sparsity of structural damage. The key of success is to construct accurate damage-reflection Lamb wave model based on the linear characteristic of simultaneous excitation system. In addition, the model performance is further improved by excitation waveform design for each transmitter. The effectiveness of the proposed method is experimentally demonstrated with array imaging of simulated delamination in composite laminates.

Analytical evaluation of resonance frequencies for planar defects: Effect of a defect shape

Local Defect Resonance (LDR) is a promising approach to detect, localize, and visualise damage based on its resonant ultrasonic response. In this paper, the classical Rayleigh's method is applied for analytical evaluation of the fundamental frequencies of LDR for various shape planar defects in polymers and composite materials. A series of experiments implemented with simulated and realistic delaminations verifies the analytical formulas obtained and reveals the constraints for practical application of the theoretical approach. The case studies considered include some basic shapes of the defects to recognize the impact of the defect shape, size and position on LDR frequency. The expressions derived combined with the resonant imaging results address the reverse problem of characterizing the depth of the defect location.

Detectability of Subsurface Defects with Different Width-to-Depth Ratios in Concrete Structures Using Pulsed Thermography

The active thermography technique is one of the most effective nondestructive tests for evaluating subsurface delaminations in concrete structures. The limitation of this method, which has been studied for some time, is that the width of the smallest detectable defect should be at least two times larger than its depth. However, controversy on this matter remains for concrete material with largely uncertain homogeneity, although the development of the infrared (IR) detector technology improved the above-mentioned limitation. In this study, the pulsed thermography (PT) technique is therefore conducted in the laboratory to investigate the detectability of delaminations with the width-to-depth ratio (w2d) ranging from 1.0 to 7.9 by using a long IR wavelength detector with a focal plane array of 640 $$\times $$ 480 pixels. The study focuses on the w2d ratio lower than 2.0. A concrete specimen was made with 12 embedded simulated delaminations having different sizes and depths. The results showed that a combination of PT and pulsed phase thermography can be used to detect delaminations with a w2d ratio equal or greater than 1.25. In addition, the absolute contrast above the delamination increases with the higher w2d ratio, indicating that even for a relatively deep delamination, it is still detectable if a delamination is provided by appropriate heat energy and its size is sufficiently large. Finally, the study also indicates that as the amount of heating energy provided is increased, the greater accuracy in predicting the depth can be obtained.