Related Experiment Video
Updated: Mar 27, 2026
09:51
A Polymer-based Piezoelectric Vibration Energy Harvester with a 3D Meshed-Core Structure
Published on: February 20, 2019
26.1K
Piezo device for micro-texturing surfaces with vision-based neural network quality assessment.
Roland Bejjani1, Cynthia AlLabaky1, Christophe Abboud1
1Department of Mechanical & Industrial Engineering, Lebanese American University, Byblos, Lebanon.
Science Progress
|March 25, 2026
Summary
A new piezo texture device creates micro-dimples on surfaces using vibrations. Artificial intelligence, specifically Convolutional Neural Networks (CNNs), analyzes these micro-features for quality control in advanced manufacturing.
Area of Science:
- Manufacturing Engineering
- Materials Science
- Artificial Intelligence
Background:
- Advanced manufacturing demands precise micro-scale feature detection and classification.
- Traditional quality control methods are insufficient for micro-features like dimples.
- Vibration-based techniques offer potential for micro-texturing.
Purpose of the Study:
- To develop a novel piezo texture device for generating micro-textures via vibrations.
- To apply artificial intelligence for real-time monitoring and analysis of imprinted micro-features.
- To classify micro-dimples and evaluate their similarity ratio using Convolutional Neural Networks (CNNs).
Main Methods:
- Development of a piezo texture device incorporating an ultrasonic concentrator and piezo-actuated sensors.
- Micro-texturing of workpieces using vibration-assisted turning.
- Implementation of GoogleNet and ResNet-50 Convolutional Neural Networks (CNNs) for image-based analysis.
- Classification of micro-dimples and assessment of their topological parameters.
Main Results:
- Successful imprinting of micro-textures (dimples) onto workpiece surfaces.
- Demonstration of CNNs' capability for accurate dimple classification.
- Evaluation of dimple similarity ratios, enabling advanced quality assessment.
- Establishment of AI as a viable alternative for micro-feature quality control.
Conclusions:
- The developed piezo texture device effectively generates micro-dimples.
- Convolutional Neural Networks (CNNs) provide a powerful tool for automated micro-feature analysis and quality control.
- This integrated approach advances capabilities in micro-scale manufacturing and inspection.
