Related Experiment Video
Updated: Jan 10, 2026
07:58
Automatic Laser-based Geometry Capture for Finite Element Analysis of Weld Beads
Published on: July 25, 2025
740
Interfacial Gap Prediction in Laser Welding of Pure Copper Overlap Joints Using Multiple Sensors
Hyeonhee Kim1,2, Cheolhee Kim3, Minjung Kang1
1Flexible Manufacturing R&D Department, Korea Institute of Industrial Technology, Incheon 21999, Republic of Korea.
Materials (Basel, Switzerland)
|November 27, 2025
Summary
This study introduces a deep learning approach for predicting interfacial gaps in copper joints using multi-sensor fusion. Spectrometer data, combined with image and optical sensors tomography (OCT), achieved 98.8% accuracy in gap classification.
Area of Science:
- Materials Science
- Manufacturing Engineering
- Artificial Intelligence
Background:
- Accurate prediction of interfacial gaps is crucial for ensuring the quality and reliability of copper overlap joints in manufacturing.
- Traditional methods for gap assessment can be time-consuming and may lack precision.
Purpose of the Study:
- To propose and validate a novel deep learning approach for predicting interfacial gaps in copper overlap joints.
- To investigate the effectiveness of multi-sensor fusion, combining image, spectrometer, and optical sensors tomography (OCT) data, for gap prediction.
Main Methods:
- Development and validation of deep learning models using data from an image sensor, a spectrometer, and OCT sensors.
- Analysis of melt pool dimensions, keyhole behavior, spectral intensity variations, and OCT-derived keyhole depth.
- Implementation of sensor fusion techniques to combine data from multiple sources.
Main Results:
- Deep learning models accurately predicted interfacial gaps based on variations in melt pool dimensions, keyhole behavior, spectral intensity, and OCT data.
- A binary gap classification model achieved a high accuracy of 98.8%.
- The spectrometer proved to be the most effective single sensor, with image and OCT sensors providing complementary information. Sensor fusion, particularly using all three sensors, yielded the best prediction performance.
Conclusions:
- Multi-sensor fusion combined with deep learning offers a powerful and accurate method for predicting interfacial gaps in copper overlap joints.
- The findings highlight the potential for improved weld quality assessment and manufacturing process optimization through advanced sensing and AI techniques.
