Abstract

The capability of microwave signals to penetrate inside composites and interact with the inner structure makes them a very attractive candidate for composite inspection. Various techniques of microwave nondestructive testing (NDT) are used for detecting disbonds in composites. Despite their promising results, these techniques suffer from poor spatial resolution due to the given features that do not significantly distinguish between the defect and defect-free regions. In this paper, a hybrid signal processing based on a refinement feature extraction method is employed to enhance the imaging efficiency of the disbond detection in composite material. This technique is based on scanning the composite material with an open-ended rectangular waveguide operating from 18 to 26.5 GHz and analyzing its reflections using the proposed hybrid signal processing method. Maximal overlap discrete wavelet packet transform is employed to provide significant informative features of each frequency point. The Bi-directional long short-term memory (Bi-LSTM) network approach is used to distinguish the significant features and the outliers. The Bi-LSTM classifies each inspected location into a defect or defect-free location. The findings presented in this paper show the advantages of the refinement method-aided microwave technique in detecting disbonds down to 1 mm with an accuracy rate of 88.84%, a significant advantage over any current disbond inspection technique.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call