Deep Learning-Assisted Fingerprint-Inspired Flexible Pressure Sensor for Tension Monitoring in Carbon Fiber...

Area of Science:

• Materials Science
• Engineering
• Sensor Technology

Background:

• Carbon fiber is crucial for aerospace, wind power, and electric vehicles.
• Controlling tension stability in wide-width carbon fiber production is a major manufacturing challenge.
• Current methods struggle to ensure consistent quality and performance in large-scale carbon fiber manufacturing.

Purpose of the Study:

• To develop a novel sensor for real-time tension monitoring of wide-width carbon fiber tow arrays.
• To improve the stability and quality of mass-produced carbon fiber.
• To introduce an intelligent system for detecting and classifying tension anomalies.

Main Methods:

• Fabrication of a flexible pressure sensor using laser etching, inspired by fingerprint structures.
• Integration of the sensor array onto a tension roller for multi-fiber tension detection.
• Development of a convolutional neural network for end-to-end tension anomaly classification.

Main Results:

• The fingerprint-inspired sensor exhibits high sensitivity (18.08 kPa⁻¹) and a wide pressure range (up to 550 kPa).
• Multi-fiber tension detection achieved a sensitivity of 5.55 N⁻¹ when the sensor array was installed on a tension roller.
• The convolutional neural network demonstrated a high tension anomaly classification accuracy of 96%.

Conclusions:

• The developed flexible pressure sensor provides a practical solution for real-time, intelligent tension monitoring in carbon fiber production.
• This technology addresses key challenges in large-scale, high-consistency carbon fiber manufacturing.
• The sensor advances sensing technology applications in flexible electronics and smart manufacturing.