Structural integrity of ultrasonically welded thermoplastic composite joints is needed
to ensure high-performance in transport, defense and infrastructure where continuous
structural health monitoring (SHM) is a key to safe and reliable operation. Defects
caused by manufacturing like voids, absence of fusion, and interfacial delamination
lead to a significant decrease in the strength of the joints and long-term stability and can
change over the course of service. This work describes the formation of a Lamb wave–
based SHM system of continuous monitoring, detection, localization, and quantification of the damage in high–density polyethylene (HDPE) lap joints. The experimental
research is combined with numerical modelling and sophisticated signal processing
to develop a predictive connection between higher and sub–harmonics characteristics
of Lamb waves and fault severity level to provide real–time condition of the joints
and pathways of the damages. Ultrasonic welding: 20 kHz (45 µm amplitude, 2.5
bar pressure, 1.0 s weld time, 1.5 s hold time) was used to prepare the joint samples.
Specimens with intentionally introduced defects (1.9 – 11.6 mm voids i.e. unfused
zone and over–melt zone) were fabricated for controlled evaluation.

The Lamb wave experiment used broadband PZT transducer as the actuator and sensor, fed with a 5 cycle Hanning–windowed tone burst at 30, 60 and 120 kHz. A Laser
Doppler Vibrometer (LDV) was used to measure the full-field wave propagation over
the weld region. The profiles of dispersion of the composite lap joint were calculated
using a semi-analytical finite element (SAFE) model and checked using frequency
wavenumber (f–k) analysis. Harmonical identification and damage–induced scattering
was analyzed using short-time fourier techniques (STFT). With the help of the finite element simulations, strong scattering of Lamb waves, lower energy transmission, and
sub-harmonic generation were found in damaged areas. The proposed SHM framework, which integrated parameters of normalized amplitude loss and nonlinear index,
facilitated quantitative forecast of strength decrease and remnant load bearing capacity
of tensile test specimens. These findings indicate that Lamb wave-based SHM offers a
quantitative, quick, and sensitive way of continuous integrity monitoring and prognostics of ultrasonically welded composite joints of thermoplastics. The method provides strong correlation with mechanical strength, excellent repeatability, and compatibility with automated SHM for smart polymer-composite manufacturing. The current study extends guided-wave-enabled quality assurance in high-performance thermoplastic structures.

Structural Health Monitoring (SHM) of critical aeronautical composite structures, such as the Airbus A380 flap track fairing, presents significant challenges. These components operate under high stress and exhibit complex geometry, resulting in inherent material anisotropy and highly variable curvature. These characteristics induce direction-dependent dispersion and complex multi-modal Lamb wave propagation, severely degrading the performance and reliability of traditional damage imaging algorithms. This paper investigates the inherent limitations of conventional Total Focusing Method (TFM) amplitude imaging and proposes a novel Phase-Frequency Coherence Imaging (PFCI) method underpinned by comprehensive anisotropic dispersion compensation. The foundation of the methodology involves accurately characterizing the anisotropic properties of the flap track fairing's surface. A rigorous process was implemented: A single step response was acquired, and the frequency response function was used to reconstruct the response across various frequencies. Group velocity-frequency curves were meticulously extracted and subjected to a segmented polynomial fitting procedure to exclude measurement errors and ensure physical validity. Numerical integration of the smoothed curves then yielded precise, angle-dependent wavenumber-frequency relations. This accurate physical model was utilized to design a dispersion-compensated excitation signal. A uniform linear piezoelectric transducer array was deployed over a 200mm x100 mm inspection area containing a 5 mm simulated circular hole. Initially, we evaluated the effectiveness of traditional amplitude-based TFM using three signal superposition strategies (coherent, incoherent, and hybrid grouping). Experimental results showed that imaging quality was severely compromised by intense direct wave interference across all three modes. Even with spatial masking, the resulting amplitude maps were dominated by array artifacts and background noise, failing to localize the damage. The failure of amplitude-based methods, which neglect instantaneous signal information, prompted the development of the PFCI approach. By establishing damage indicators based on phase consistency and frequency consistency, this method is designed to simultaneously suppress dispersion effects and mitigate the coherent noise originating from direct waves and structural internal reflections. The resulting phase-frequency damage images demonstrate superior damage identification capability and remarkable robustness against the complex Lamb wave propagation paths found in anisotropic flap track fairings. This novel approach significantly advances the potential for reliable defect detection in high-value, geometrically complex aeronautical structures.

Guided ultrasonic waves (GUW) are widely used in structural health monitoring due to their large propagation range and high sensitivity to impedance changes. In particular, B-scan representations enable a detailed observation of wave–damage interactions. However, the quantitative evaluation of Lamb wave modes (A0 and S0) is significantly affected by interference from overlapping wave components, leading to reduced measurement accuracy, especially when weak damage-induced signals are present.

This study investigates and compares two fundamentally different filtering approaches for improving signal separability in x–t wavefields: a data-driven image-processing method based on directional derivatives and a physics-based method using two-dimensional Fourier transformation (2D-FFT). The directional derivative filter enhances wave components aligned with a chosen orientation by emphasizing spatial gradients, while the 2D-FFT projects the wavefield into the frequency-wavenumber (f-k) domain, enabling selective filtering of individual modes and propagation directions followed by inverse transformation.

Both methods are applied to experimental B-scan data of Lamb wave interactions with barely visible impact damage (BVID). Their performance is quantitatively evaluated using a signal-to-noise ratio (SNR) to distinguish between measurement noise and wave modes.

The results show that directional derivative filtering improves visibility only when the target wave is already distinguishable in the raw data, providing limited or no SNR enhancement in strongly interfered regions. In contrast, the 2D-FFT-based approach significantly increases SNR across all evaluated regions by effectively separating wave modes and propagation directions in the f-k domain. This enables reliable extraction of weak damage-related features that are otherwise not observable.

Overall, the study demonstrates that 2D-FFT filtering is superior for the quantitative analysis of weak Lamb wave interactions in the presence of modal interference, making it a robust tool for advanced damage characterization in structural health monitoring applications.

Pipes made of glass fiber reinforced plastic (GFRP) are used in a wide range of applications. The combination of high strength and rigidity with low weight and excellent corrosion and chemical resistance makes them superior to conventional plastic or metal pipes in many areas of application. They are characterized by their chemical resistance to a wide range of acids and salts, lack of electrical conductivity and good tightness. Due to their properties, they are primarily used in drinking water and sewage pipes, in the chemical industry, as pressure pipes and downhole tubing and casings in geothermal applications.

Matrix cracking and, to a lesser extent, delamination are considered to be the main reason for the failure of GFRP pipes. This involves the detachment of laminate layers or the liner from the structural composite, which can lead to leaks or reduced load-bearing capacity and sudden failure. Matrix cracking and delamination are primarily caused by existing operating loads. GFRP pipes are primarily tested using visual inspection, thermography, ultrasound or acoustic emission testing. Apart from acoustic emission testing, it is not possible to test or monitor large areas. Therefore, guided waves are another promising testing method where waves propagate along the pipe and are guided by the inner and outer surfaces. Compared to ultrasonic waves, guided waves are less strongly attenuated and can propagate several meters through pipes which enable monitoring large areas.

In this paper, guided waves were used to distinguish between matrix cracked/delaminated and intact pipe segments. During a pressure test, pipes were pressurized to burst, and matrix cracked/delaminated areas occurred as a result. Measurements were then taken on five damaged pipe segments and, with five intact pipes serving as a control group. Statistical methods were implemented to identify the damaged pipes and to distinguish from the control group of intact pipes. A network of piezoelectric sensors on the pipes was used to send and receive guided waves. Repeated application of the network to the pipes caused deliberate variability in the measurement data, which is used to evaluate the fault detection probability of the method. To compare the measurements with a model, the dispersion curve of the GFRP pipes was calculated in advance and the recorded measurement data was verified against it. With an adapted model for the material parameters, the combined damage mechanisms in the pipes are also simulated using the dispersion curve.

The use of fiber-reinforced polymer composites is becoming increasingly important across many industries. To meet growing demands, numerous molding and manufacturing processes are being developed, including a particularly cost-effective technique: vacuum resin infusion. However, this process still requires optimization in many respects, especially in terms of monitoring the material’s state during the curing phase of the composite.

This work aims to propose a method for monitoring the degree of polymerization (α) of the resin during its curing phase by tracking its rheological properties in real time using non-intrusive techniques, specifically Lamb waves emitted and received by a network of piezoelectric transducers. The choice of Lamb waves is particularly adapted to the VRI process, as it is mostly used for plate and shell parts.

The proposed approach is as follows:
(i) Construction of a wave propagation model for the mold–resin system, approximated as an isotropic solid plate coated with a visco-elastic fluid layer on its upper side.
(ii) Simulation of the waves that can propagate through the system depending on the rheological and geometrical properties of the viscoelastic layer.
(iii) Comparison of the simulations with full-field experimental measurements performed using a laser vibrometer during the resin curing process, in order to validate the theoretical model.
(iv) Monitoring of a full curing process using a PZT network under an aluminium plate, aiming to validate the link found numerically between rheological properties of the fluid layer and attenuation of Lamb waves.

The model described above is supported by a preliminary reactional and rheological study of the epoxy resin under various temperature conditions, in order to monitor the evolution of dispersion laws directly as a function of α. These elements make it possible to infer the mechanical properties of the viscoelastic layer from the ultrasound measurements.
Full field experimental results using the laser vibrometer showed a strong agreement between the measurements and the numerical prediction, letting the propagation model be used as a strong and accurate virtual testing tool. A series of parametric studies were conducted using the validated model, centred around the evolution of the thickness of the fluid and its viscosity and shear modulus. Experiments using calibration silicon oils and an epoxy resin during its curing phase were later conducted in order to assess the accuracy of the model for a wide range of use.

Ultimately, this study and the developed method aim to provide an in situ and non-intrusive monitoring technique for the whole process from its resin filling phase to the end of the curing reaction, thereby facilitating the widespread adoption of composite manufacturing across different industrial scales while ensuring quality and efficiency standards are properly met.