Related Experiment Video
Updated: Jun 23, 2026
08:31
Three-dimensional Optical-resolution Photoacoustic Microscopy
Published on: May 3, 2011
18.2K
Stable 3D Deep Convolutional Autoencoder Method for Ultrasonic Testing of Defects in Polymer Composites.
Yi Liu1, Qing Yu1, Kaixin Liu2
1Institute of Process Equipment and Control Engineering, Zhejiang University of Technology, Hangzhou 310023, China.
Polymers
|June 19, 2024
Summary
A new 3D deep convolutional autoencoder (3D-DCA) accurately detects defects in polymer composites. This method overcomes noise and echo interference, enabling precise defect size, shape, and depth determination.
Area of Science:
- Materials Science
- Non-destructive Testing
- Artificial Intelligence
Background:
- Ultrasonic testing is crucial for polymer composite defect detection due to its speed and reliability.
- Challenges exist in accurately identifying defects in ultrasound images due to echo interference and noise.
Purpose of the Study:
- To develop a robust method for enhanced defect detection in polymer composites.
- To address limitations of traditional ultrasonic testing in interpreting defect signals.
Main Methods:
- A stable three-dimensional deep convolutional autoencoder (3D-DCA) was developed.
- 3D convolutional operations were used to learn spatiotemporal data properties.
- A dual-layer encoder and depth receptive field (RF) were implemented to mitigate echo effects.
Main Results:
- The 3D-DCA method effectively identified defects in polymer composites.
- The approach accurately determined defect size, shape, and depth.
- Mitigation of surface and bottom echo interference was achieved.
Conclusions:
- The developed 3D-DCA offers a reliable solution for defect detection in polymer composites.
- This method improves the accuracy and interpretability of ultrasonic testing data.
- Feasibility demonstrated on carbon-fiber-reinforced polymers.
