RESUMEN

This article presents a method to use the dispersive behavior of ultrasonic guided waves and neural networks to determine the isotropic elastic constants of plate-like structures through dispersion images. Therefore, two different architectures are compared: one using convolutions and transfer learning based on the EfficientNetB7 and a Vision Transformer-like approach. To accomplish this, simulated and measured dispersion images are generated, where the first is applied to design, train, and validate and the second to test the neural networks. During the training of the neural networks, distinct data augmentation layers are employed to introduce artifacts appearing in measurement data into the simulated data. The neural networks can extrapolate from simulated to measured data using these layers. The trained neural networks are assessed using dispersion images from seven known material samples. Multiple variations of the measured dispersion images are tested to guarantee the prediction stability. The study demonstrates that neural networks can learn to predict the isotropic elastic constants from measured dispersion images using only simulated dispersion images for training and validation without needing an initial guess or manual feature extraction, independent of the measurement setup. Furthermore, the suitability of the different architectures for generating information from dispersion images in general and an image-to-regression visualisation technique, are discussed.

RESUMEN

Adhesively bonded composite joints can develop voids and porosity during fabrication, leading to stress concentration and a reduced load-carrying capacity. Hence, adhesive porosity analysis during the fabrication is crucial to ensure the required quality and reliability. Ultrasonic-guided wave (UGW)-based techniques without advanced signal processing often provide low-resolution imaging and can be ineffective for detecting small-size defects. This article proposes a damage imaging process for adhesive porosity analysis of bonded composite plates using UGWs measured by scanning laser Doppler vibrometer (LDV). To implement this approach, a piezoelectric transducer is mounted on the composite joint specimen to generate UGWs, which are measured over a densely sampled area. The signals obtained from the scan are processed using the proposed signal processing in different domains. Through the utilization of filter banks in frequency and wavenumber domains, along with the root-mean-square calculation of filtered signals, damage images of the adhesive region are obtained. It has been observed that different filters provide information related to different void sizes. Combining all the images reconstructed by filters, a final image is obtained which contains damages of various sizes. The images obtained by the proposed method are verified by radiography results and the porosity analysis is presented. The results indicate that the proposed methodology can detect the pores with the smallest detectable pore area of 2.41 mm2, corresponding to a radius of 0.88 mm, with an overall tendency to overestimate the pore size by an average of 11%.

RESUMEN

To assess the ability of structural health monitoring (SHM) systems, a variety of prerequisites and contributing factors have to be taken into account. Within this publication, this variety is analyzed for actively introduced guided wave-based SHM systems. For these systems, it is not possible to analyze their performance without taking into account their structure and their applied system parameters. Therefore, interdependencies of performance assessment are displayed in an SHM pyramid based on the structure and its monitoring requirements. Factors influencing the quality, capability and reliability of the monitoring system are given and put into relation with state-of-the-art performance analysis in a non-destructive evaluation. While some aspects are similar and can be treated in similar ways, others, such as location, environmental condition and structural dependency, demand novel solutions. Using an open-access data set from the Open Guided Waves platform, a detailed method description and analysis of path-based performance assessment is presented.The adopted approach clearly begs the question about the decision framework, as the threshold affects the reliability of the system. In addition, the findings show the effect of the propagation path according to the damage position. Indeed, the distance of damage directly affects the system performance. Otherwise, the propagation direction does not alter the potentiality of the detection approach despite the anisotropy of composites. Nonetheless, the finite waveguide makes it necessary to look at the whole paths, as singular phenomena associated with the reflections may appear. Numerical investigation helps to clarify the centrality of wave mechanics and the necessity to take sensor position into account as an influencing factor. Starting from the findings achieved, all the issues are discussed, and potential future steps are outlined.

Asunto(s)

RESUMEN

Data-driven analysis for damage assessment has a large potential in structural health monitoring (SHM) systems, where sensors are permanently attached to the structure, enabling continuous and frequent measurements. In this contribution, we propose a machine learning (ML) approach for automated damage detection, based on an ML toolbox for industrial condition monitoring. The toolbox combines multiple complementary algorithms for feature extraction and selection and automatically chooses the best combination of methods for the dataset at hand. Here, this toolbox is applied to a guided wave-based SHM dataset for varying temperatures and damage locations, which is freely available on the Open Guided Waves platform. A classification rate of 96.2% is achieved, demonstrating reliable and automated damage detection. Moreover, the ability of the ML model to identify a damaged structure at untrained damage locations and temperatures is demonstrated.

Asunto(s)

RESUMEN

In many industrial sectors, structural health monitoring (SHM) is considered as an addition to nondestructive testing (NDT) that can reduce maintenance effort during the lifetime of a technical facility, structural component, or vehicle. A large number of SHM methods are based on ultrasonic waves, whose properties change depending on structural health. However, the wide application of SHM systems is limited due to the lack of suitable methods to assess their reliability. The evaluation of the system performance usually refers to the determination of the probability of detection (POD) of a test procedure. Up until now, only a few limited methods exist to evaluate the POD of SHM systems, which prevents them from being standardized and widely accepted in the industry. The biggest hurdle concerning the POD calculation is the large number of samples needed. A POD analysis requires data from numerous identical structures with integrated SHM systems. Each structure is then damaged at different locations and with various degrees of severity. All of these are connected to high costs. Therefore, one possible way to tackle this problem is to perform computer-aided investigations. In this work, the POD assessment procedure established in NDT according to the Berens model is adapted to guided wave-based SHM systems. The approach implemented here is based on solely computer-aided investigations. After efficient modeling of wave propagation phenomena across an automotive component made of a carbon-fiber-reinforced composite, the POD curves are extracted. Finally, the novel concept of a POD map is introduced to look into the effect of damage position on system reliability.

Asunto(s)