Related Experiment Video
Updated: Jul 2, 2025
11:34
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
Published on: May 15, 2017
11.1K
Quantitative Detection of Defects in Multi-Layer Lightweight Composite Structures Using THz-TDS Based on a
Dandan Zhang1,2,3, Lulu Li1,2,3, Jiyang Zhang1,2,3
1Key Laboratory of Optoelectronic Measurement and Control and Optical Information Transmission Technology, Ministry of Education, Changchun University of Science and Technology, Changchun 130022, China.
Materials (Basel, Switzerland)
|February 24, 2024
Summary
A novel U-Net-BiLSTM network accurately identifies defects in multi-layer composites using terahertz non-destructive testing. This advanced method achieves high accuracy for defect detection and quantitative analysis in aerospace materials.
Area of Science:
- Materials Science
- Aerospace Engineering
- Non-Destructive Testing
Background:
- Multi-layer lightweight composites are crucial in aviation and aerospace.
- Manufacturing and usage can introduce defects, necessitating effective non-destructive testing (NDT).
- Terahertz (THz) NDT faces challenges in defect recognition due to complex structures and signal variability.
Purpose of the Study:
- To develop an advanced model for accurate defect identification and quantitative analysis in multi-layer composite structures.
- To address the limitations of existing methods in THz non-destructive testing of complex materials.
- To improve the reliability and precision of defect detection in aerospace components.
Main Methods:
- Introduction of a hybrid U-Net-BiLSTM network combining spatial and temporal feature extraction.
- Optimization of network architecture and parameters for THz spectroscopy data.
- Application of the model for identification, classification, and quantitative analysis of defects using THz NDT.
Main Results:
- The U-Net-BiLSTM network achieved 99.45% accuracy and 99.43% F1 score in defect identification.
- The proposed model outperformed existing networks like CNN, ResNet, U-Net, and BiLSTM.
- Successful reconstruction of 3D THz defect images through defect classification and thickness recognition.
Conclusions:
- The U-Net-BiLSTM network offers a highly effective solution for quantitative defect detection in multi-layer composites via THz NDT.
- This approach significantly enhances the accuracy and reliability of NDT for aerospace applications.
- The study demonstrates the potential of deep learning for advanced material inspection and quality control.
