Defect severity classification of complex composites using CWT and CNN

Document Type

Article

Publication Date

1-1-2022

Abstract

Composite structures are prone to internal defects such as delamination. Due to this, it is vital to recognize internal flaws in composite materials accurately because there is possibility that these internal defects can severely degrade the composite structure’s strength. This work aims to develop an intelligent complex composite defect severity classification which will contribute to efficient monitoring of composite structures during their service life. Firstly, the behavior of guided ultrasonic waves is processed and transformed into image database using continuous wavelet transform method. Then, a defect classification framework is proposed by using convolutional neural network to classify six types of defect sizes. A total of 798, 342, and 90 images are used for training, validation, and testing, respectively. The results present that the proposed system achieved approximately above 86 of precision and recall for all six defects classes. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Keywords

Complex networks, Composite structures, Convolutional neural networks, Deep learning, Defects, Glass ceramics, Structure (composition), Complex composites, Composite defects, Composites material, Composites structures, Deep learning, Defect severity classification, Internal defects, Internal flaws, Signal classification, Wavelets transform, Wavelet transforms

Divisions

sch_ecs

Funders

None

Publication Title

Lecture Notes in Electrical Engineering

Volume

834

Publisher

Springer Science

Additional Information

Cited by: 0; Conference name: International Conference on Computational Intelligence in Machine Learning, ICCIML 2021; Conference date: 1 June 2021 through 2 June 2021; Conference code: 274369

This document is currently not available here.

Share

COinS