In Guided Wave Structural Health Monitoring (GW-SHM), structures are instrumented by sparse transducers arrays to detect defects through the tiny change induced by the presence of a defect on the GW propagation within the structure. Due to the complexity of the GW propagation, various post-processing techniques have been proposed to analyze the measurements, among which guided wave imaging (GWI) allows both defect detection and localization at the same time. Defect quantification is also possible with post processing techniques such as guided wave tomography; however, such approaches require a large number of sensors and a good knowledge of the specimen to achieve the quantification, hence are not applicable in uncontrolled environment and if the number of sensors is limited. Moreover, defect sizing based on image processing techniques is not directly possible from GWI as the images depend on a wide variety of parameters such as transducer positions and number, wavelength of inspection and even defect position.

This communication presents the characterization of defects based on guided wave imaging acquired with a sparse transducer array on a composite panel with an impact-caused delamination using deep learning. Delay and Sum imaging is used to provide a physical representation of the structure’s health. Convolutional Neural Networks (CNN) are chosen to interpret the full images provided by guided wave imaging without dimensionality reduction to automate detection, localization and sizing tasks. CNN are trained with custom-made databases generated by the SHM module of CIVA. Results show good capacity to invert real defects using the numerically trained neural network.


Defect Characterization in Composites With Deep Learning in Guided Wave Shm

No proceedings papers in 2020. Video only.

QNDE2020 - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.